AI and Data Driving India’s Energy Transformation for Climate Solutions
20 Feb 2026 10:00h - 11:00h
AI and Data Driving India’s Energy Transformation for Climate Solutions
Session at a glance
Summary
This discussion focused on building a global data and AI ecosystem to address climate and energy challenges, particularly in India, through Data.org’s ClimateVerse initiative. Dr. Cormekki Whitley opened by explaining how Data.org operates capacity accelerators across five regions to build AI practitioners while helping impact-focused organizations adopt these tools responsibly. The initiative aims to unlock climate and energy data through local talent development and digital transformation, addressing current barriers like fragmented ecosystems, lack of standardized data formats, and limited access to hyper-local information.
Two detailed case studies demonstrated practical applications of this approach. Arthur Global presented research on heat’s impact across Delhi neighborhoods, surveying 27,500 Indians and finding that 45% reported household illness from heat-related issues. Their neighborhood-level analysis revealed significant temperature variations based on green cover and urban planning, with productivity losses of 50% during heat waves and doubled energy consumption for cooling. ClimateDot showcased their work standardizing India’s power sector data, which suffers from inconsistent nomenclature and non-interoperable formats across states and departments. They developed automated systems to aggregate disparate data sources into machine-readable formats, supporting the India Energy Stack initiative.
The expert panel discussion emphasized critical enabling conditions for scaling these solutions beyond pilot projects. Key themes included the need for coordination at scale across multiple stakeholders, standardized data architectures, and cross-functional capacity building. Panelists stressed the importance of defining success metrics clearly, ensuring inclusive design processes, and developing AI literacy among policymakers and practitioners. The discussion concluded with recognition that moving from innovation to institutional adoption requires sustained collaboration between academia, government, industry, and civil society organizations.
Keypoints
Major Discussion Points:
– Building Climate and Energy Data Infrastructure: The discussion focused on creating unified, standardized data architectures for climate and energy sectors, particularly in India. Speakers emphasized the need to move from fragmented, non-interoperable data systems to machine-readable, standardized formats that can support AI applications and cross-sector collaboration.
– Moving from Pilots to Systemic Change: A central theme was the challenge of scaling innovative climate-AI solutions beyond pilot projects to achieve institutional adoption and sustained impact. This includes embedding data-driven tools into core organizational and government decision-making processes rather than keeping them as isolated innovations.
– Coordination and Collaboration at Scale: Multiple speakers highlighted the critical need for coordination across stakeholders – from government agencies and regulators to private sector and civil society. The “whole of ecosystem approach” was emphasized as essential for managing complex, multi-stakeholder climate and energy initiatives affecting billions of people.
– Capacity Building and Cross-functional Skills: The discussion emphasized developing “socio-technical skills” – creating professionals who are bilingual in both domain expertise (climate, energy, health) and data/AI capabilities. This includes building AI literacy among policymakers, NGOs, and industry leaders to enable better decision-making about AI implementations.
– Equity and Just Transition: Speakers addressed the importance of ensuring climate and energy transitions are inclusive, particularly for vulnerable populations like workers transitioning from fossil fuel industries. The discussion emphasized that technological solutions must consider social equity and support those at the “bottom of the pyramid.”
Overall Purpose:
The discussion aimed to explore how to accelerate climate and energy data ecosystems for sustained public impact, moving beyond individual innovations to create systemic change. The session sought to identify enabling conditions, governance structures, and capacity-building needs necessary to institutionalize climate-AI solutions at scale.
Overall Tone:
The tone was collaborative and solution-oriented throughout, with speakers building on each other’s insights rather than debating. It maintained a balance between technical expertise and practical implementation challenges, with an underlying sense of urgency about climate action. The discussion was forward-looking and optimistic about technology’s potential while remaining realistic about institutional and coordination challenges. The tone remained consistently professional and constructive, with speakers sharing concrete examples and actionable frameworks.
Speakers
Speakers from the provided list:
– Dr. Cormekki Whitley – Works at Data.org as a connector, convener, and catalyst; involved in capacity accelerator network (CAN) building global workforce for data and AI practitioners
– Professor Neelanjan Sircar – Director of the Centre for Rapid Insights at Arthur Global; focuses on providing policy relevant feedback in a rigorous but timely manner
– Karan Shah – Chief Operating Officer of the India Office of Arthur Global; works with governments, philanthropists, multinationals and other policy stakeholders to improve policy design and implementation
– Dr. Srikant K. Panigrai – Director General, Indian Institute of Sustainable Development and Distinguished Research Fellow; policymaker working on scientific policies for 37 years; former global climate negotiator for India
– Srinivas Krishnaswamy – Works at Vasudha Foundation; involved in creating the India Climate and Energy Dashboard adopted by NITI Aayog
– Priyank Hirani – Director of Capacity Building at Data.org
– Swetha Ravi Kumar – Head of FSR Global; currently leading the India Energy Stack Program
– Akhilesh Magal – Works at ClimateDot; focuses on organizing India’s power sector data and building unified, scalable data architecture
– Speaker 1 – Role/expertise not clearly specified in the transcript
– Dr. Priya Donti – Assistant Professor at MIT working on developing AI for power grid optimization and renewables integration; Co-founder of Climate Change AI nonprofit
Additional speakers:
None identified beyond the provided speakers names list.
Full session report
This comprehensive discussion explored the critical challenge of building robust climate and energy data ecosystems to drive sustained public impact, particularly in India, through Data.org’s ClimateVerse initiative. The session brought together researchers, policymakers, and practitioners to examine how data-driven climate solutions can move beyond pilot projects to achieve systemic transformation and institutional adoption.
The ClimateVerse Vision and Global Context
Dr. Cormekki Whitley opened the session by positioning Data.org as a connector, convener, and catalyst operating five data capacity accelerators across the U.S., India, Latin America, Africa, and the Asia Pacific. This global network aims to build a workforce of data and AI practitioners whilst helping impact-focused organisations unlock these tools responsibly. The ClimateVerse initiative emerged from recognition that climate and energy domains—including health, energy, productivity, and livelihoods—are fundamentally interconnected, requiring integrated approaches rather than siloed solutions.
The initiative’s discovery work revealed persistent barriers to effective climate action: fragmented ecosystems, lack of shared language and standards, and limited access to hyper-local information, particularly in emerging economies. Through over 50 consultations and reviews of more than 40 data platforms and tools in India alone, the team identified that data and tools must become more discoverable, granular, and interoperable, supported by appropriate incentives and infrastructure, and paired with interdisciplinary capacity building and stronger multi-stakeholder collaboration.
Heat as a Macroeconomic Variable: The Delhi Case Study
Arthur Global’s research, presented by Karan Shah and Professor Neelanjan Sircar, fundamentally reframed heat from a meteorological phenomenon to a macroeconomic variable. Their comprehensive study surveyed 27,500 Indians across 20-plus states and 490-plus assembly constituencies, revealing startling impacts: 45% of respondents reported household illness from heat-related issues in the previous month, with two-thirds experiencing symptoms for more than five days.
The research identified a critical scale mismatch between current heat action plans, typically designed at state or district levels, and the neighbourhood-level experience of heat. Their Delhi-focused study of 2,400 households demonstrated that heat experience depends on three key parameters: individual characteristics (occupation, daily routines, economic background), neighbourhood design (planning, density, tree cover), and geographic location (temperature, humidity, airflow patterns).
The findings revealed dramatic spatial variations in heat experience across Delhi. Areas with better spatial planning and green cover showed temperature differences of up to one degree Celsius compared to densely populated areas with limited green space. Increasing green cover from 4% to 10% resulted in one degree of cooling—a significant impact during heat waves. The economic implications were substantial: a three-degree Celsius increase in experienced heat led to 50% increases in work loss, whilst people using air conditioning reported three times better sleep but consumed twice as much energy.
Crucially, the research revealed that cooling has become a private adaptation strategy, with over 40% of comfortable respondents relying on air conditioners or coolers, highlighting the absence of adequate public adaptation responses. This finding underscores systemic inequality in climate resilience, where individuals bear the cost and responsibility of heat adaptation whilst public infrastructure responses remain inadequate.
Standardising India’s Power Sector Data Architecture
Akhilesh Magal from ClimateDart addressed the challenge of India’s power sector data, which, whilst significant and granular, remains largely unstructured and non-interoperable. The organisation has spent three to four years developing solutions to organise this data at state levels, building learnings that can scale nationally.
The problems are multifaceted: inconsistent nomenclature (such as “O&M” versus “Operations and Maintenance” in different years), varying data granularity (where detailed categories suddenly become lumped together), and non-interoperable systems across different government portals. These inconsistencies create significant barriers for machine learning applications, requiring substantial human intervention to standardise data before AI tools can process it effectively.
ClimateDart’s response involved building intelligent scripts that scout the internet, scrape relevant data, and aggregate it into standardised, machine-readable formats. Whilst acknowledging this approach’s inefficiency compared to API access, they’ve created a unified data acquisition method and architecture for the power sector. Their work with Goa state demonstrates the potential: a 15-year historical database accessible through interactive dashboards (available via QR codes) that track critical metrics like renewable power obligations.
The ultimate vision connects to the India Energy Stack initiative—creating digital public infrastructure for India’s energy sector analogous to UPI’s transformation of banking. This would enable unprecedented possibilities, such as rooftop solar owners in Tamil Nadu selling electricity to consumers in Ladakh, unlocking new economic opportunities through standardised data architecture.
Expert Panel: Enabling Conditions for Systemic Change
The expert panel discussion, moderated by Priyank Hirani, focused on the critical institutional shifts needed to move from innovation to sustained adoption. Due to time constraints, speakers provided focused insights on their key areas of expertise.
Coordination and Governance at Scale
Shweta Ravi Kumar from FSR Global introduced the AAA framework (Architecture, Adoption, Accelerator) for coordination at scale. The Architecture component encompasses technical specifications and standards that enable common data language across systems. Adoption recognises that stakeholders operate at different capability levels, requiring diverse pathways for engagement. The Accelerator creates sandbox environments for building use cases that demonstrate value extraction for different stakeholders.
The India Energy Stack program, involving the Ministry of Power and regulatory bodies, exemplifies this whole-of-ecosystem approach, moving beyond government coordination to include all relevant stakeholders in the design process. Recent demonstrations showed concrete examples like farmer “Arun from Meerut” selling power to “garment owner Lakshmi in Delhi” through simple WhatsApp interfaces, illustrating how complex AI systems can be made accessible through familiar user interfaces whilst maintaining technical sophistication in the background.
Data Quality and Institutional Barriers
Srinivas Krishnaswamy from Vasudha Foundation shared insights from developing the India Climate and Energy Dashboard, now adopted by NITI Aayog. The dashboard aggregates data from multiple ministry reports and platforms, providing a unified view of India’s power and energy sector trends viewed through climate and development lenses. With 5 lakh users (500,000) from 170 countries and an average of 2,000 hits per day (reaching as much as 5,000 hits per day), the platform demonstrates significant global demand for integrated climate and energy data.
However, persistent challenges remain: continued reliance on manual data entry despite digital integration possibilities, reluctance to share even non-sensitive data, and sluggishness in data access that prevents real-time integration. These institutional barriers highlight the need for stronger coordination between multiple data-collecting agencies at national and state levels, along with more granular data collection at higher frequencies.
Cross-Functional Capacity Building and AI Literacy
Dr. Priya Donti from MIT emphasised the critical need for AI literacy at scale among policymakers, NGOs, and industry leaders. She noted that despite widespread AI discussions, very few decision-makers can actually define AI or understand AI pipelines, creating barriers to effective implementation and governance. This knowledge gap trickles down to organisational and policy decisions about AI integration.
The discussion highlighted the need for “socio-technical skills”—professionals who are bilingual in both domain expertise (climate, energy, health) and data/AI capabilities. This requires moving beyond the current dichotomy between building capabilities in-house versus external procurement, towards developing specialised solutions providers that understand the nuanced aspects of specific sectors rather than offering generic AI solutions.
Equity and Just Transition
Dr. Srikant K. Panigrai, Director General of the Indian Institute of Sustainable Development and Distinguished Research Fellow, brought attention to equity considerations in climate and energy transitions. As India rapidly expands renewable energy capacity, workers transitioning from coal-based industries face livelihood security challenges without adequate training in renewable energy technologies.
Drawing on Gandhi’s talisman of considering the impact on the most vulnerable, Dr. Srikant emphasised that technological solutions must ensure “nobody is left behind” and support those at the “bottom of the pyramid.” He referenced his institute’s work on apiculture research projects that study honey bee behavior and pollination while providing livelihood sources for tribal women, demonstrating how climate research can directly benefit vulnerable communities.
Defining Success and Measuring Progress
A critical theme throughout the discussion was the importance of defining success metrics and solution frameworks from the outset. Dr. Priya Donti argued that many initiatives fail to scale because they build without clear objectives or measurement frameworks, making it difficult to know whether intermediate successes are leading towards final goals.
The conversation revealed the need for principled approaches to defining what solutions are and what success means, including understanding the role of technical systems versus human decision-makers. This requires establishing clear metrics that are stated and measured, along with intermediate success milestones that create pathways for scaling.
Key Challenges and Ongoing Initiatives
The session identified several ongoing initiatives while acknowledging significant challenges that remain. The India Energy Stack program continues developing national data policy frameworks for the power sector, whilst the unified data architecture approach demonstrated in Goa will expand to other states.
However, substantial challenges persist: overcoming institutional reluctance in data sharing, scaling manual processes to automated systems with API integration, developing specialised solutions providers for specific sectors, and creating sustainable funding and governance models for long-term data infrastructure maintenance.
The discussion highlighted the need for tiered approaches that balance data openness with security requirements, hybrid capacity building that combines in-house development with external expertise, and gradual transition programs for workers moving between energy sectors.
Moving Toward Systemic Transformation
This exploration of climate and energy data ecosystems reveals both the tremendous potential and significant challenges in scaling data-driven climate solutions. The session demonstrated that technical capabilities exist to address critical climate challenges, from neighbourhood-level heat management to cross-state energy trading. However, realising this potential requires fundamental shifts in how institutions coordinate, share data, and build capacity.
The emphasis on moving from pilots to permanent solutions reflects a broader maturation in the climate-tech space, where the focus is shifting from proving technical feasibility to achieving sustained institutional adoption. The strong consensus among diverse stakeholders on core challenges—data fragmentation, need for standardisation, importance of cross-functional skills, and requirements for inclusive design—provides a foundation for coordinated action.
The session ultimately demonstrated that successful climate and AI integration requires treating technology as part of broader socio-technical systems that must account for human behaviour, institutional dynamics, and equity considerations. The vision of coordination at scale, exemplified by initiatives like the India Energy Stack, offers a pathway for transforming how societies respond to climate challenges through integrated data and AI capabilities while ensuring that benefits reach all segments of society.
Session transcript
Data .org is a connector, a convener, and a catalyst. Through five data capacity accelerators in the U.S., India, Latin America, Africa, and the Asia Pacific, our capacity accelerator network, or CAN, is building a global workforce for data and AI practitioners. While helping impact -first organizations unlock these tools in service of their missions, through CAN, we invest both in supply and demand, strengthening the pipeline and advancing the readiness of organizations to think, plan, and operate responsibly in an AI -driven world. Our work is globally informed and locally grounded through more than 100 cross -sector partners. In India, we focus on climate and its deep implications. We have many intersections with help, energy, productivity, and livelihoods. These domains may appear distinct, but they are fundamentally interconnected.
That insight on intersectionality gave rise to ClimateVerse while we’re here today. ClimateVerse, a vision to unlock climate and energy data, tools, and collaboration pathways by upskilling local talent and supporting digital transformation for organizations. Let me share a bit about what we’ve learned about the climate and energy data ecosystems during our discovery work. Reliable, usable data is essential for decision -making and policy. But today, many barriers persist. Fragmented ecosystems. Lack of shared language and standards. And a lack of accessible, hyper -local information, especially in emerging economies. In India alone, we conducted 50 -plus consultations. We reviewed 40 -plus data platforms and tools. So we’ve been talking to a whole lot of people and listening to a whole lot of people and learned alongside CAN partners like Junhagra, Civic Data Lab, and SEAS, amongst others.
What we heard was that data and tools must be easier to discover, more granular, interoperable, and supported by incentives and infrastructure, and paired with interdisciplinary capacity building and stronger multi -stakeholder collaboration. So it’s the listening and the hearing and joining. India is already doing important work in this space, but the real questions now are, how do we move from pilot… to system level change? How do we design ecosystems that drive adoption, not just innovation? And how do we build the interdisciplinary talent that can translate across climate and AI? To integrate climate and energy data into real decision -making, we need to build local capacity and advance organizational AI readiness and activate partnerships across academia, practitioners, industry, and government.
We all have a role to play. Today we want to share examples of what we’ve been building with our partners and invite all of you alongside our expert panelists, which you will see and hear from later, to help identify the gaps, the enablers, and the conditions needed to drive impact at scale for climate resilience and a global clean energy transition. Transition. With that, let me invite our first partner from Arthur Global for our first Climate Solutions Spotlight, Dr. Linan and Karan Shah, to share insights from their recent study on spatializing the impact of heat on human health and productivity across Delhi’s neighborhood with implications for grid planning. Welcome.
Okay. Thank you very much, Cormekki, and very good morning to all of you who are here today. Thank you for being there. At the outset, I need to thank our wonderful, lovely partners, Data .org and the entire team for not only facilitating the event but facilitating the study that we’re going to present today. My name is Karan. I’m the Chief Operating Officer of the India Office of Arthur Global. We’re a policy organization that works with governments, philanthropists, multinationals and other policy stakeholders to improve the design and implementation of policy making. I’m here with my colleague Neelanjan Sircar Sarkar who’s the director of the Centre for Rapid Insights which is our rapid insights unit that aims to support governments and partners with providing policy relevant feedback in a rigorous but timely manner.
So with that I just like to talk a little bit about our work that we recently did. So we know that being in Delhi extreme heat is no longer episodic, it is a structural phenomena that we’re dealing with. We’re not talking about heat waves as shocks anymore, we’re talking about a significant rise in the baseline. When Delhi records its warmest night in six years we know something is going wrong. There is no relief, nights are no longer providing that relief anymore. And the invisible part of all this is not the temperature, right? The invisible part is the impact on health burden, productivity, and grid management, right? Today, we know that 76 % of our population actually lives in districts that are classified as high to very high heat risk, and close to 50 % of India’s population actually works in the outdoors.
So if India needs to think about its productivity and competitiveness, and cities are going to be the engines of economic growth, and cities are going to be dependent on labor markets, then we know that heat no longer is just a meteorological variable, but is now a significantly important macroeconomic variable. So our work on heat actually has been going on for several years. So back in 2024, between the months of May and June, ARSA actually conducted… India’s largest survey to try and integrate the impact of heat on the health of citizens. We surveyed 27 ,500 Indians across 20 plus states and about 490 plus assembly constituencies to try and discover three things. What is the impact of heat on health and how are citizens coping both at home as well as their workplace?
The results, as you will see, are startling. Close to 45 % of respondents actually reported to have one member of their household ill in the last one month because of a heat -induced issue. And close to two -thirds of those actually felt sick for more than five days. Now you can just sort of try and understand the impact on productivity here. And when you start digging into the data, you realize that heat has very, very uneven disturbances, actually impacting the less privileged population. Significantly more, right? Even coping gave us a lot of insights. So greater than 30 % of people actually said that they are uncomfortable in their own home. and even from the ones that said that they are comfortable, more than 40 % relied on either air conditioners or coolers.
Now this tells us that cooling has become a private adaptation strategy. We still don’t have a public one. So that was the motivation of our study and what made it very clear that heat has very, very widespread impact and that impact is not evenly distributed. So we said, okay, how is heat distributed then? And we looked at cities as a critical part to identify that. Now we all know about the urban heat island effects in cities. Cities amplify heat, distribute it even more unevenly. Concretized areas are causing heat traps. Building materials are actually keeping heat much longer. The lack of adequate tree cover is causing natural ventilation and natural cooling to actually disappear. We know all of these things are actually impacting heat very, very much.
Now here as well, we found that that our response architecture is failing, right? Most heat action plans in the country today are made either at the state level or the district level. But heat is significantly experienced at the neighborhood level, right? And that’s the scale mismatch that we wanted to highlight with the study to try and see if heat action plans can be more granularly informed, right? Now, we began our hypothesis just on three parameters. And we said the way in which heat is experienced actually rests on three parameters. The first parameter is who you are, right? What’s your occupation? What are your daily routines? What appliances do you own, right? What sort of economic background do you belong to?
We said that has a significant impact on the way you will get exposed to heat as well as deal with heat. So that was the most important contribution to the study, is to bring the voice of citizens and layer that with other forms of data. The second question we asked is, how is your neighborhood built? And this is not your district or your city, this is your immediate neighborhood. Is it well planned, is it formal, is it informal, is it dense, does it have a lot of tree cover, does it not have enough tree cover? Those are the aspects that we looked at. And third is where you live. So even where you live actually makes a big difference because temperature, humidity, pockets of airflow and ventilation can make a substantial difference and cause pockets of uneven heat across cities.
So the hypothesis was that these are the three pillars on which we will be able to understand the impact of heat on households. And that’s what led to the study. So with that, I’d just like to welcome Professor Neelan to walk us through what some of these findings were and talk about what implications does this have on plans as well as grid management.
so just taking over from that great introduction from my colleague Karan so let me just talk you through what the data problem here is because that’s a large part of what we’re here so we have good data from satellites on green cover, on built area we have good measures from the Indian Meteorological Department on air temperature land temperature, humidity what we don’t have is the third piece of the puzzle which is how are people experiencing heat we know that experiencing heat has a substantial amount to do with behavior do you have an air conditioner do you work in the heat do you have comorbidities these are pieces of information that you need to be able to triangulate with these other administrative data sets now if this data does not exist in any system in a systematic way then how do you make claims about health heat action plans, energy overload, right?
You need this piece of data. So our empirical problem was the following. If I go to a person’s household, right? I need to be able to construct the built environment for that person, I need to construct what kind of heat that person’s experiencing, but I also need to construct what that person is doing throughout the day, right? I need to know whether that person’s turning on the air conditioner, at what time, I need to know when that person is working, where that person is working. So that’s where the surveys come into place. Now our infrastructure at the Center for Rapid Insights basically uses that geographic information, that spatial information, figures out where to sample, and in this case we sampled 2 ,400 households broadly across the city of Delhi, and collect that data very quickly, because heat waves don’t last for very long, so we did this all in two weeks, right?
So that’s the kind of technology that one needs to be able to do with data collection. Just very quickly going through some of the results. You can see that there are huge differences between when an area is more spatially planned and not. This difference is about a degree, right? So if you happen to live on the right side where there’s more green space, you are experiencing a degree less of heat in the middle of a heat wave than somebody living in a more densely populated area. This is the area right around the airport, so many of us will be coming in and out of this area. This is just a snapshot of what’s happening there, where you can see that a large part of this story is actually the amount of green cover.
If I just increase the green cover by 5 to 6 percentage points from 4 % to 10%, we’re talking a degree of cooling. We also wanted to demonstrate that actually heat and how people are experiencing heat have very, very significant economic impacts on productivity. So you can see that there’s a 50 % increase in work loss in the middle of a heat wave. Just for a 3 degree Celsius increase in experience heat, right? So this is actually not uncommon. If you look back at some of these initial maps, you can see it’s going from 39 to 46. so actually the variation is 7 to 8 degrees of Celsius in terms of what people are feeling in Delhi and just 3 degrees Celsius is increasing work loss by 50 % so we’re talking about very very significant economic productivity effects so how are people coping with this kind of heat well it turns out and this is something that exists in literature more generally beyond cooling that exists in the environment and across much of India you have environments that look densely densely concretized like what we have on the left without green cover, people are having to turn on their ACs people do report being having 3 times better sleep if they’re turning on the air conditioning but they also report consuming twice as much energy so as the world gets hotter if people are going to require turning on the air conditioning to get better sleep to be able to show up to work the next day we know it’s going to have an impact on the grid.
And I just want to make one quick point here. Without doing this kind of measurement I might be able to look at energy flows over the last two years and guess what the next month of grid load will look like. But it’s going to be very hard to predict three years down the line, five years down the line unless you know who’s using an AC, how much they’re using the AC. So that kind of grid load management is what’s important. So just finishing up here. So what I want to demonstrate here and I think what we want to demonstrate at ARCA Global individual characteristics, built environment characteristics are so determinative of how people experience heat that without very localized heat action plans that integrate all of this data we can’t really get to people and address their needs.
The other thing is when it comes to grid planning yes I might be able to plan for the electricity grid tomorrow or maybe a year down the line but if I need to have planning for 5 years down the line, 10 years down the line without this kind of data, how individuals are using air conditioners, when they’re using air conditioners, how they’re cooling how the world is changing for them, you won’t be able to come up with adequate grid planning. Thank you.
Thank you Karan and Neelan for those great insights Next up we would like to share another example of a use case in the AI and energy space from ClimateDot and I invite Akhilesh Magal to talk about their work on open data architecture and how it will shape multiple use cases for India’s energy stack Thank you
Thank you All right, good morning, ladies and gentlemen. Am I audible? Yes? Okay. It’s great to be here. Thanks to data .org, who we’re working with extensively on reshaping some of energy power sector data, actually, in India. And it’s also nice to see familiar faces in the auditorium. So I think this is going to be a short but sweet, hopefully sweet presentation. Happy to interact with some of you if you have some questions after this. All right, so what we’ve been doing as ClimateDart is trying to get a grip on India’s power sector data, which is significant, large, and often disorganized. We have data. We have granular data. The issue is, of course, getting it into usable formats.
And so over the last three or four years, we’ve been, as an organization, trying to organize some of this data. We’ve been trying to get some of this data at the state level, and trying to build learnings that can be scaled up to the national level. and I’ll talk about some of the collaborations that we have in this regard. So what is the problem? As I said, India’s power sector, we have a lot of data, significant number of data points, but it’s largely unstructured and non -interoperable. And this is a problem especially when we want to talk to each other between states, for instance, or between the center and the states, but also within the states.
We’ve noticed discrepancies between years. For example, you’ll see on the right side, you’ll see two rather simple examples. We have many more, but given the paucity of time, I’m focusing on two examples. You’ll see that, for example, in the first table, you’ll see O &M being the acronym being used, but on the right side, you see in the earlier year, in 2016, we have it being fully the expanded version of that. Now, that may seem a very small issue for us as humans, but when you have machines reading this, you already have the first stumbling block, and that would require… I think it’s a very big issue. I think it’s a very big issue. significant man hours or woman hours or people hours in terms of, to be accurate, to be in order to make sure that the machines read this and we can, you know, build AI tools and so on on top of this.
So one of the problems is on data nomenclature, but we also have problems in data granularity. And what does that mean? For example, those of you in the power sector will recognize these terms. Fixed charge and variable charges have standard reporting metrics for the power sector. And in 2022, we had that data, so it’s pretty granular. But in 2023, we noticed from the regulatory filings, this has suddenly disappeared and it’s been lumped into a single cost head. Now, for people working in the power sector, this may be okay, and we may be able to do some simple math and get these numbers out. But for machines, this is already a significant problem. So as we built out the databases, standardized databases, we realized that these are some of the problems that we could already begin to share with regulators.
With regulators, with policy makers, with data scientists, et cetera, so that we begin to organize this. already. And so what we’ve worked on for the last two and a half years also with support from our partners, data .org, our funders and so on, we’ve tried to build a unified and scalable data architecture for India’s power sector that works across states and within states as well. And I’ll tell you why the within states is so important. So what we want to do is we want to get the data from various plethora of input sources. We have PDFs and scanned reports. Sometimes these are handwritten reports in government files that have been digitized or scanned, often from a mobile phone.
So you need to use some sort of character recognition, some basic form of intelligence to be able to read that. And we’ve run into significant challenges there. Most of the other data is on spreadsheets and databases, easier to read. But the challenge is that these aren’t really organized in the way we would like them to be organized. So they’re not consistent. of course now the government has worked significantly in putting out a lot of data in the public domain, in portals and so on, each department having their own portals, but the problem is most of these portals don’t really talk to each other so they are, not just the front end is different but also the back end is very different, so smiling so I know this is a problem and of course we have significant number of data silos that we just don’t know how to access sometimes for good reasons because these are isn’t data that you can make public but sometimes this is publicly available data that is sitting in silos, so can we begin to have a discussion on making this accessible so what we’ve done is really over the last 3 or 4 years, built scripts intelligent scripts that can sort of scout the internet verse, get this data that we want, scrape the data and aggregate this data, this is not efficient some of you have a computer science background, you know that this is typically not a very efficient way to do this, what would be efficient is to have an API, API access to this and so what we’ve done with all the scraping is built a standardized data acquisition method but also an architecture for the power sector the key point in the outcome is to make this standardized and machine readable what this means is if we can get this data read by machines with very little human interaction that’s the best because it really increases the pace at which we can begin to bring various state related data onto a single homogenized architecture we of course can the applications from this are multi, there’s a multiverse of what we can do with these applications what we’ve been doing is building analytical dashboards so power sector dashboards at the state level and I’ll show you on the next slide what we’ve done but we can also build AI insights so any AI engine today requires machine readable data, data is extremely important so once we have these databases which various tools can plug in and building AI tools on top of this becomes really, really easy.
I mentioned the API aspect and I think that’s critical but then all of this can go into making better policies and effective decision making which is what we do as an organization. So a small example of what we did for the state of Goa. Goa, we’re working with the state to understand how to bring up bring in all the power sector data into a single portal. This is 15 year data goes back in history and that’s just an example on the right side of one of the pages of that portal where we were tracking their renewable power obligation something very important especially from a climate and an energy transition perspective. The QR codes are up there so if some of you are interested you can scan the QR codes it should take you directly to the website and it’s a very interactive very visually built dashboard.
And I have one minute left so I’m… just wrapping up. Thanks. so essentially walking through this process of automation, standardization and visualization so we need automation we need to reduce manual intervention we need to standardize a lot of this which we believe we’ve done for at least two or three states and of course then build interesting tools that are usable not just tools that look at past data but perhaps modeling tools and predictive tools that look at what the past sector might be in the next five years so extremely crucial from a policy perspective so that leads us to the India energy stack and I’ll talk very little about this because there are people who are leading this initiative Shweta, my colleague is here so it’s an initiative built or rather led by the Ministry of Power the RAC and FSR Global Shweta is here it’s essentially the digital public infrastructure for India’s energy sector very similar to the UPI country this but I is for power and I is for power what UPI is for banking in India, unlocked a trillion dollar, two trillion dollar economy something like this, so can we do something similar for the Indian power sector where someone from Tamil Nadu can sell electricity from their rooftop power plant to someone in Ladakh, if this is possible I think our work as researchers is really would come to fruition so we can certainly take questions on this in our panel discussion but I will wrap up here and I thank you very much for your attention Thank
Thank you so much for that Thank you so much for that, you’ve heard some great presentations about what’s possible with data, but remember the data is about the people at the end of the day so there are many more such climate and AI solutions that innovators in the room will be able to share but for the next segment of this session I want to invite my colleague Priyank Hirani, Director of Capacity Building at data .org to explore the enabling conditions to accelerate the climate and energy data ecosystem for sustained public impact with an esteemed panel of global experts. Priyank.
Thank you, Cormekki, and thank you to our wonderful speakers. We’re going to be quick on this one. We’re running out of time. But I quickly want to bring on key experts so that my talking is minimal on this panel and you get a chance to listen to these global visionaries. So let me first invite Dr. Srikanta Panigraha. So please join us. Mr. Srinivas from Vasudha Foundation, Dr. Priya Donte from MIT, and Shweta Ravikumar from FSR Global. Thank you so much. So, today’s panel is going to focus on not just technology, but thinking about the enabling conditions. We heard about two use cases, and I’m sure a lot of you in this room are working on climate and AI use cases, and you have several examples.
But as Kurmiki mentioned in an opening remark, sort of how do we move from pilots to permanents? How do we move from having just dashboards to ensuring decisions and sustained decisions with those things? And how do you help these innovations to be institutionalized? So that’s the goal for us to cover in the next 25 to 30 minutes. We want to think about these enabling conditions, whether they are on the governance side, what are the incentives, what are the digital public infrastructure needed, what sort of coordination mechanisms that might be needed, and most importantly, what’s the capacity within organizations and as a country that we need to develop? So what’s the talent pipeline that we need to think about?
So we’re going to start with the talent pipeline. and how do essentially we start measuring these things both quantitatively and qualitatively so that we are able to track the progress. So with that, let’s begin with the big picture. My first question that I’m going to ask all the panelists to quickly reflect on is from your vantage point, what is the single most critical institutional shift or enabling condition that might be needed to ensure that these solutions become embedded in both the core organizational or say government decision making rather than remaining as one of innovations. So maybe we’ll go around in this order. Srinivas. And please feel free to quickly introduce yourself or tell us about your organization.
is incredible. So we need to leverage that and that can be leveraged if we have the data. Now in terms of institutional and governance I would say that let’s take India today. We have multiple agencies that have been tasked in compiling and collecting the data. At the national level you have the Bureau of Energy Efficiency that compiles data on all efficiency related aspects. You have the Central Electricity Authority. You have the Ministry of Statistics and Planning Implementation. At the state level you have the State Planning Board. So on and so forth. But then what is still lacking is the granular data collection and compilation. That’s something that is still lacking I would say. And so that’s where institutions need to gear up to ensure that we have more granular collection and compilation of data at a higher frequency of sharing.
So that’s how I would put that.
Thank you so much. That’s very insightful. Dr. Srikanth, what’s one critical institutional shift that you think is needed?
I am Dr. Srikanth K. Panigrahi, Director General, Indian Institute of Sustainable Development and Distinguished Research Fellow. I am basically a policymaker working on scientific policies because of my interest for last 37 years. Now I am leading this institute which is a public policy think tank and scientific research organization, Indian Institute of Sustainable Development. So coming to the questions, in public policy when you are insurable to people and you are insurable to planet and you are insurable to the growth of the nation, sustainability rests in all three of them. You need, you have to be very particular. that analysis -based decision -making has to be adopted. And analysis -based decision -making is only possible when you are adopting tools, scientific tools like AI is a wonderful tool and which has the precision and it helps you with the exact information and the data you are looking for.
If wrong data will be fed to the tool, the wrong decisions will be indicated. So as it has been told by my colleague, we need so far is quality of the data, relevance of the data and all these has to be in alignment with the objective which we are looking for. And so for that we need for the right public policy, we need right data strategy and there are many examples. to which I am not getting into. And in IAST, we have a wonderful research project where we are studying apiculture. That is the behavior of bees, honey bees. And these bees are generating, through pollinisation, they are generating honey, which is a good livelihood source for the poor tribal women.
So I will explain this study in my later round.
Thank you, sir. So I’m hearing sort of ensuring coordination between departments, ensuring thinking about the data strategy. What more do you have to add, Swetha?
Thanks, Priyank. I’m Swetha, head of FSR Global, currently leading the India Energy Stack Program. So I’m going to share some learnings from there. You used the word coordination, and that’s literally on every slide that I have on IES, which is coordination at scale. We’re talking about designing systems for… Billionaires. so I think the government has already started to take its steps in terms of this whole of a government approach what we have done through this initiative is taken that to a whole of an ecosystem approach because we need in such a multi -sector multi -stakeholder projects we need all of them at the design board if we don’t articulate what is in it for me for every stakeholder from early on the question that you asked can we move from pilots to scale would be a recurring question so to have them in the drawing board is very important and in terms of actually scaling in the AI unlock I think inclusivity is a very important aspect we need to consider Akhilesh was just talking about can I trade from Tamil Nadu to another place in fact two days ago in this very room we facilitated that trade and showed how a farmer Arun from Meerut was selling to a garment owner in Lakshmi in Delhi across state borders and they did it through very simple WhatsApp based interfaces because they didn’t want to understand all of this complicated AI.
That’s for all of us engineers who love to work with complicated things. As a consumer, they could talk in their local language to an AI bot in WhatsApp and trade power. It needs to be made as simple as that for the stakeholders. Ultimately, all of the best ideas in this room need to scale in countries like ours and beyond.
Got it. Thank you. I like the phrase coordination at scale, thinking about the billions. Dr. Priya.
Hi, everyone. I’m Priya Donti. I am an assistant professor at MIT working on developing AI for power grid optimization and renewables integration. I’m also a co -founder of Climate Change AI, which is a nonprofit focused on large -scale democratization and coordination of skills and expertise in AI and climate. I agree with everything the other panelists have said. The two things I will add. One is being principled about defining what success means. The other is being principled about defining what solutions are. I think often we’re building without doing that. And it leads to things like when we have, let’s say, a pilot innovation, we don’t know where we’re headed. We don’t know what that intermediate success is leading to a final success since we don’t set up stages for actually moving things forward.
We also kind of defining what success means means having metrics that are kind of stated that are measured. It means thinking about what is the role of the technical system versus the human who’s making a decision around it. So I think basically kind of anchoring in that notion of what is success, how do we measure it, how do we get there, what is an intermediate success, I think drives a lot of really important thinking and infrastructure around this. The second thing I would say is having, I think we’ve heard a lot about quality. Coordination, but I think also being principled about what kind of cross -functional skills are necessary to actualize and measure solutions in the long term and kind of what that means in terms of gaps, in terms of what kinds of actors exist in the broader ecosystem to make that happen.
Right now, there’s a little bit of a dichotomy where kind of. you know between kind of can we build capabilities in -house versus can we procure externally and when it comes to external procurement often there’s some sort of generic notion of there’s like some notion of a solutions provider that does generic data that does generic AI and yet in many places kind of solutions are very specific right we heard about power system related like data standardization that kind of effort is really important but it also looks very different if you’re doing that in health if you’re doing that in buildings and if you don’t have kind of specialized solutions providers that are really able to contend with this the kind of nuanced aspects of knowing the data and knowing the methods in a particular domain then I think there’s often sort of a gap where there isn’t enough capacity to upskill internally nor is there actually a good procurement option so I think from a public policy perspective kind of you know I think there’s often sort of a gap where there isn’t enough capacity to upskill internally nor is there actually a good procurement option so I think from a public policy perspective kind of enabling this more diverse ecosystem of solutions providers that are also more tuned towards the needs of specific sectors is also important
that’s wonderful and that’s core to sort of the philosophy of data .org that we think about where it’s essentially putting people at the center of the problem and so thank you for rounding us up Priya because as all the things I was hearing it’s ultimately about do we have the skills do we have the institutional capacity to be able to engage with these things and that is something that we need to look at from the cross functional skilling perspective the lens that you talked about and that’s what at data .org we often talk about as socio -technical skills so how do we think of people as bilinguals in terms of domain understanding but also a data or AI understanding and then they are able to work across.
So continuing with that thought I want to come back to Swetha and Swetha you talked about IES and the coordination that you’re doing with sort of multiple kinds of stakeholders bringing everyone together from a regulatory and governance perspective thinking about this ecosystem of the energy sector What ecosystem design choices, it could be sort of standards, interoperability, it could be things around incentives. Do you think most influence whether stakeholders meaningfully adopt data -driven tools? The one thing that you talked about, which I really love, is ensuring that they are there at the table from the get -go. They’re not an afterthought. No one wants to be an afterthought. But amongst these things of standards, interoperability, incentives, what do you think ensures sustained adoption of tools?
Thank you. I’m going to break it down through what we call the AAA framework at the Indie Advertising Stack. So first is the architecture, which is all of the technical specifications. I’m not using the word standards because standards means it’s an authoritative stamp, right? So it’s a combination of standards and specifications and new things coming where the old cannot sort of adapt to. So it’s going to be a suite of specifications and standards that allow for all of us to have sort of a common data language. Let’s put it that way. so that if you and I want to exchange information, we know what and how to do that. If two systems need to do as we saw in the use cases, they know how to do that.
And the power sector is quite complicated. You have millions of assets and millions of people interacting, so we need to have a basket of solutions that need to come together and be interoperable at the core. Then, of course, the second one is the adoption because not all of us are at the same level playing field as stakeholders. There are some DISCOMs who have certain systems built in, some ready to build in, which might be an advantage in their case because you can leapfrog. You don’t have to think about integrating into legacy systems. So we’ll have to create these different pathways for different stakeholders to harness this data AI layer or digitalization wave that’s coming about in the sector.
And that’s being done through what we call in the accelerator, the third A, wherein you’re building use cases so that everyone can plug and see what value extraction that they can have. And some descoms might want to focus on grid phasing use cases. Some might want to look at market side. Some might want to look at societal impact. So there has to be this pieces of puzzle that could sort of fit in for each of them. And it’s not something that you do over a year and close. It’s a continuous process of building. And so through the accelerator, which is a sandbox environment, we’re building certain reference implementation architectures demonstrating the idea into action. And then it’s for the ecosystem to take and scale with the stakeholders.
And that’s where I said the articulation of what is in it for me. And that’s where incentives come in. and we also have the regulators on board co -designing with us and the policy makers in parallel. Ministry of Power is bringing in a new national data policy framework for the power sector because we’re talking about critical infrastructure here. We need to also look at who gets to access what kind of data and what should be sort of the safeguards we have within the ecosystem. So it’s truly a 360 -degree view on this particular project and hopefully we will have some best practices out of this, learnings from here that could help other projects.
Got it. Thank you. I love the AAA framework. We’re going to keep coming back to it. I wanted to bring in Srinivas into the conversation now and your work at Vasudha over so many years has supported the NITI IO through the India Climate and Energy Dashboard, which is now adopted and institutionalized. So you’ve seen sort of this coordination piece, getting everyone aboard. Getting the adoption sustained. that the full cycle in practice, apart from all the other work that all of you do with the state government. So from this experience, like what strengths did you find in India’s climate and digital architecture while working on that dashboard or working with the government? And what and I’d be remiss not to ask, like what gaps do you think currently are preventing further coordinated action?
So I would I would start off by saying that the data that is there in the India climate and energy dashboard is not new. It’s there and multiple reports of various ministries and agencies. It is there in multiple dashboards of various ministries and agencies. But what the ICD does, it brings together data from all these various reports, all the various dashboards in one unified manner. And it actually marries. This is from the entire power sector and energy and power sector value chain. And it marries the data with the climate. Data and key economic indicators. so what it actually does is it gives you a holistic picture of what are the trends and developments in India’s power sector, power and energy sector but viewed from a climate and a development lens so that’s what the ICED does second what it does is that the visual architecture in a way has been designed that it brings out the nuances of the trends so it’s not just about aesthetics, yes we did take care of aesthetics, we did want good looking graphs but we also wanted graphs and infographics that brings out the key nuances that one is looking at to give a holistic picture of what is happening in this entire sector if you are looking at energy transition you can actually get what are the trends now if you look at the kind of users of the ICED from a low of about average of 2000 hits per day we get as much as 5000 hits per day with roughly how many I would say 5 lakh users across multiple stakeholder groups and from 170 countries, so virtually the entire world.
Okay, we have 195 countries, so we have 170 countries, we have hits from 170 countries. Now, that’s the kind of impact that the ICD has had. So it’s not just in India, but it’s also global. Coming to the second point on the challenges. I think the biggest challenge that we still have today is that we still have dedicated staff who have to do manual entry of the data. I think in today’s time and age, I think we should have digital integration. We should have APIs that Akhilesh talked about. I think that is something that is still lacking. Yes, for some of the data sets, we are able to digitally scrap it. But then by and large, we have, and Rahul is here, and you can see we have a dedicated team who are just into this manual entry.
Thank you. and that’s a pain because not only does it mean that errors tend to seep in and we have to do a lot of quality checks but also means that the ICT still remains near real time and we want to make it real time. So now we have a 3 to 4 days gap but we would like to ideally make it real time. The third, the second challenge I would say is that there is still a reluctance even for non -sensitive data to share the data. I would say a combination of reluctance and a combination of sluggishness. Sometimes when you are dealing with getting the data it’s like pushing a wet sponge. It’s as sluggish as that and that sometimes gets a little tricky because we are very conscious that we want to have this as a real time and so when the sluggishness seeps in then things tend to get a little slow.
I would like to add one other point. now if you look at how do we avoid duplication of efforts in which I think I would one thing that we at Vasudha have always been endeavouring and not just with ICD but all the dashboards that we created with states whether it’s a Gujarat Climate Action Tracker, Tamil Nadu Tracker, whether it’s a Kerala Dashboard or even the predecessors of the ICD which was Vasudha Power .in or Vasudha EMI one thing that we made very clear is that the data is available in open domain. Anybody can use it there are no paywalls. The whole idea was to reduce duplication of efforts and also ensure that people can share the data.
Thank you so much. I think that idea of reducing barriers to access and making any tool user friendly is super critical. I want to bring Dr. Shrikanta into the conversation and think about the aspects of equity, just transition, long term resilience. From your experience, you’ve been a key global climate negotiator for India, you’ve been part of the IOC for many years. What operational governance and human capacity factors do you think most enable and ensure not just technically robust solutions are integrated, but also then are leading to those decisions within those systems?
A very important question indeed. When in the public policy, the equity is extremely important. And equity means the entire planning has to be inclusive. Like in UN SDGs, we have a slogan, nobody should be left behind. We have to carry everyone along with us. So the Gandhi, the Gandhi, the Gandhi, the Gandhi, the Gandhi, the Gandhi, the Gandhi, these talismans also tell the same thing. So coming to the very fundamental, the kind of the energy transition that is taking off. India is doing excellent in enhancing its renewable energy capacity, which is geometrically increasing. Say it’s solar, we are getting into wind, or say it’s other new form of energy, renewable energy, like in Ladakh, geothermal, wave energy, there is a huge investment, new projects are coming up.
So India is considered as one of the most serious nation who is heavily investing in renewables. And trying to make the transition rapid. And if you see our achievements, it is also very impressive so far. Coming to the fact, when someone is switching from coal -based fossil energy to renewable energy, the kind of the workers, the technology, everybody goes through a transition. And… And for a country like India, where the use of machine is less, and more and more people like laborers and wage -based laborers, they work at the bottom of the pyramid. Those who are engaged in coal -based work, they don’t have alternative. They are not trained in renewable energy space. So, they are very much afraid of losing their job and livelihood security.
Coming to the electric vehicles also, in mobility transition, the similar challenges are aired. In ISD, we have a separate transition research cell, where both the mobility as well as energy transition, while happening, how the transition can be taken up. of enabling the bottom line of the pyramid for giving them right training and capacity building and bringing to the mainstream of livelihood, ensuring their security has been assured. For all these things, technology plays a very big role and we need to plan and do this with precision, with optimization of time and a very focused strategic approach. The program has to be initiated for tools like different tools of AI is of great importance. Given the time, I would like to explain our B project.
which is extremely impressive, which we are taking up with National Anusandan Research Foundation and this project ensures the the pollination rate of the bee enhances more and more honey is collected from the flowers and gives better livelihood option to the poor tribal women and more honey you cannot collect unless there is more greenery so for that the more plantation densification of forest and agriculture enabling carbon credits through sequestration thank you
on that note I wanted to bring in Dr. Priya in thinking about how do we build this workforce at scale how do we get the collaboration between these different practitioners
absolutely and I will keep my remarks brief I realize we need to wrap up and so I guess the one thing I will say is that it is incredibly important that we really think about AI literacy at much larger scale among kind of policymakers, NGOs, industry, so forth. We’re having a whole AI summit, and I think the number of people who could actually define what AI is and what an AI pipeline looks like is extremely small. And this trickles down in many ways because then decision makers who are making decisions about AI at an organizational level, at a policymaker, it’s very hard to pinpoint what’s actually needed if you don’t have that basic literacy. So I will make a plug.
Climate Change AI is running an open registration virtual summer school towards the end of this year, kind of focused on trying to provide some of these AI basics as well as climate basics to those coming from an AI background to try to spur collaboration. So whether through that or something else, I would just encourage everyone, take a couple of hours to take AI 101.
Got it. Thank you so much. Thanks, everyone. Thank you so much to our panelists, and thank you for being here. We’ll pass it on to the next session. Thank you. Thank you. Thank you.
Dr. Cormekki Whitley
Speech speed
120 words per minute
Speech length
666 words
Speech time
332 seconds
Fragmented ecosystems and lack of shared standards
Explanation
The current data ecosystem is broken into isolated silos and there is no common language or standards, which prevents decision‑makers from accessing reliable, interoperable data. This fragmentation hampers the ability to translate data into actionable climate policies.
Evidence
“Fragmented ecosystems.” [1]. “Lack of shared language and standards.” [2].
Major discussion point
Data ecosystem challenges and need for granular, interoperable data
Topics
Data governance | Artificial intelligence | Capacity development
Moving from pilots to system‑level change
Explanation
Scaling climate‑AI solutions requires moving beyond isolated pilots to system‑wide adoption, which in turn depends on building interdisciplinary talent that can bridge climate science and AI. This talent is essential for sustained impact across sectors.
Evidence
“India is already doing important work in this space, but the real questions now are, how do we move from pilot… to system level change?” [78]. “And how do we build the interdisciplinary talent that can translate across climate and AI?” [68].
Major discussion point
Translating data into policy, scaling solutions, and institutional shifts
Topics
Capacity development | Artificial intelligence | Enabling environment for digital development
CAN invests in supply and demand to build AI‑ready workforce
Explanation
The Capacity Accelerator Network (CAN) funds both the supply of skilled data/AI professionals and the demand from organizations, creating a pipeline that strengthens AI readiness for climate and energy challenges.
Evidence
“…through CAN, we invest both in supply and demand, strengthening the pipeline and advancing the readiness of organizations to think, plan, and operate responsibly in an AI‑driven world.” [101]. “…our capacity accelerator network, or CAN, is building a global workforce for data and AI practitioners.” [119].
Major discussion point
Building capacity and talent pipeline for climate‑AI integration
Topics
Capacity development | Artificial intelligence | Enabling environment for digital development
Karan Shah
Speech speed
162 words per minute
Speech length
1024 words
Speech time
378 seconds
Heat impacts vary at neighborhood level, need hyper‑local data
Explanation
Heat exposure is highly heterogeneous within cities, with neighborhoods experiencing different temperature and humidity conditions. Effective mitigation therefore requires data at a hyper‑local scale.
Evidence
“But heat is significantly experienced at the neighborhood level, right?” [16]. “So even where you live actually makes a big difference because temperature, humidity, pockets of airflow and ventilation can make a substantial difference and cause pockets of uneven heat across cities.” [28].
Major discussion point
Data ecosystem challenges and need for granular, interoperable data
Topics
Environmental impacts | Data governance | Social and economic development
Heat‑action plans must be neighborhood‑specific
Explanation
Current heat‑action plans are drafted at state or district levels, which is too coarse to address the uneven heat exposure observed at the neighborhood scale. More granular plans are needed to target interventions effectively.
Evidence
“Most heat action plans in the country today are made either at the state level or the district level.” [21]. “And that’s the scale mismatch that we wanted to highlight with the study to try and see if heat action plans can be more granularly informed, right?” [87].
Major discussion point
Translating data into policy, scaling solutions, and institutional shifts
Topics
Environmental impacts | Policy & governance | Data governance
Survey shows health burden, productivity loss, uneven exposure
Explanation
A large city‑wide survey revealed that heat imposes a hidden health and productivity burden, with impacts varying sharply across neighborhoods, highlighting the need for targeted interventions.
Evidence
“The invisible part is the impact on health burden, productivity, and grid management, right?” [30]. “So that was the motivation of our study and what made it very clear that heat has very, very widespread impact and that impact is not evenly distributed.” [29].
Major discussion point
Case studies illustrating impact
Topics
Environmental impacts | Social and economic development | Data governance
Professor Neelanjan Sircar
Speech speed
177 words per minute
Speech length
954 words
Speech time
323 seconds
Absence of household‑experience data limits health and grid forecasts
Explanation
While satellite and meteorological data exist, there is no systematic collection of how households experience heat, which is essential for linking climate to health outcomes and grid load projections.
Evidence
“what we don’t have is the third piece of the puzzle which is how are people experiencing heat… if this data does not exist in any system in a systematic way then how do you make claims about health heat action plans, energy overload, right?” [26]. “Accurate household data essential for long‑term grid planning” [31].
Major discussion point
Data ecosystem challenges and need for granular, interoperable data
Topics
Data governance | Environmental impacts | Capacity development
Rapid, spatially resolved household surveys needed
Explanation
The team conducted a rapid, city‑wide household survey of 2,400 homes in two weeks, demonstrating that timely, granular data collection is feasible and critical during short heat waves.
Evidence
“we sampled 2,400 households broadly across the city of Delhi, and collect that data very quickly, because heat waves don’t last for very long, so we did this all in two weeks, right?” [27].
Major discussion point
Case studies illustrating impact
Topics
Monitoring and measurement | Data governance | Capacity development
Accurate household data essential for long‑term grid planning
Explanation
Future grid planning over 5‑10 years requires detailed information on household cooling behavior; without it, planners cannot reliably forecast load or design resilient infrastructure.
Evidence
“if I need to have planning for 5 years down the line without this kind of data… you won’t be able to come up with adequate grid planning.” [31]. “Without doing this kind of measurement I might be able to look at energy flows over the last two years and guess what the next month of grid load will look like.” [32].
Major discussion point
Translating data into policy, scaling solutions, and institutional shifts
Topics
Data governance | Environmental impacts | Monitoring and measurement
Akhilesh Magal
Speech speed
174 words per minute
Speech length
1531 words
Speech time
526 seconds
Inconsistent nomenclature and granularity block machine‑readability
Explanation
Data sets suffer from varied naming conventions and coarse granularity, creating a stumbling block for automated processing and AI tools that require machine‑readable inputs.
Evidence
“So one of the problems is on data nomenclature, but we also have problems in data granularity.” [39]. “Now, that may seem a very small issue for us as humans, but when you have machines reading this, you already have the first stumbling block…” [40]. “But for machines, this is already a significant problem.” [47].
Major discussion point
Data ecosystem challenges and need for granular, interoperable data
Topics
Data governance | Artificial intelligence | Capacity development
Standardized, API‑driven data enables AI tools for policy
Explanation
Transforming large, unstructured power‑sector data into standardized, API‑accessible formats makes it machine‑readable, allowing AI models to generate policy‑relevant insights.
Evidence
“we have a lot of data… but it’s largely unstructured and non‑interoperable.” [37]. “I mentioned the API aspect and I think that’s critical but then all of this can go into making better policies and effective decision making…” [99].
Major discussion point
Translating data into policy, scaling solutions, and institutional shifts
Topics
Artificial intelligence | Data governance | Environmental impacts
Power‑sector data fragmentation hampers analysis; API and dashboards create usable, machine‑readable datasets
Explanation
The power sector’s data is scattered across portals and formats; building scripts, APIs, and dashboards converts this fragmented information into a unified, machine‑readable architecture that supports AI‑driven analysis.
Evidence
“what we’ve done with all the scraping is built a standardized data acquisition method but also an architecture for the power sector … to make this standardized and machine readable … AI engine today requires machine readable data.” [97]. “I mentioned the API aspect and I think that’s critical…” [99].
Major discussion point
Case studies illustrating impact
Topics
Data governance | Artificial intelligence | Environmental impacts
Srinivas Krishnaswamy
Speech speed
159 words per minute
Speech length
709 words
Speech time
266 seconds
Institutional gap in high‑frequency granular data collection
Explanation
Current institutions do not collect or share data at the necessary frequency and granularity, limiting real‑time climate‑energy analysis.
Evidence
“But then what is still lacking is the granular data collection and compilation.” [9]. “And so that’s where institutions need to gear up to ensure that we have more granular collection and compilation of data at a higher frequency of sharing.” [48].
Major discussion point
Data ecosystem challenges and need for granular, interoperable data
Topics
Data governance | Monitoring and measurement | Capacity development
Coordination among multiple agencies needed for granular data sharing
Explanation
Several agencies are tasked with data collection, but without coordinated mechanisms the resulting datasets remain siloed and under‑utilized.
Evidence
“We have multiple agencies that have been tasked in compiling and collecting the data.” [60]. “So we need to leverage that and that can be leveraged if we have the data.” [59].
Major discussion point
Translating data into policy, scaling solutions, and institutional shifts
Topics
Data governance | Enabling environment for digital development
Climate‑Energy Dashboard relies on manual data entry, limiting real‑time insight
Explanation
The dashboard aggregates data from many reports, but manual entry creates a 3‑4‑day lag, preventing real‑time monitoring of climate‑energy indicators.
Evidence
“I think the biggest challenge that we still have today is that we still have dedicated staff who have to do manual entry of the data.” [152]. “So now we have a 3 to 4 days gap but we would like to ideally make it real time.” [54].
Major discussion point
Case studies illustrating impact
Topics
Monitoring and measurement | Data governance | Environmental impacts
Dr. Srikanth K. Panigrahi
Speech speed
Default speed
Speech length
Default length
Speech time
Default duration
Data quality and relevance must align with policy objectives
Explanation
For data‑driven tools to support climate policy, the data must be both high‑quality and directly relevant to the specific policy goals being pursued.
Evidence
“we need so far is quality of the data, relevance of the data and all these has to be in alignment with the objective which we are looking for.” [55].
Major discussion point
Data ecosystem challenges and need for granular, interoperable data
Topics
Data governance | Capacity development | Monitoring and measurement
Adoption of analysis‑based decision‑making requires coherent data strategy
Explanation
Decision‑makers must adopt analysis‑based approaches, which depend on a clear data strategy that ensures the right data feeds into AI tools for accurate insights.
Evidence
“that analysis -based decision -making has to be adopted.” [13]. “And so for that we need for the right public policy, we need right data strategy and there are many examples.” [56].
Major discussion point
Translating data into policy, scaling solutions, and institutional shifts
Topics
Artificial intelligence | Data governance | Capacity development
AI literacy for policymakers, NGOs, and industry is critical
Explanation
Without basic AI literacy, policymakers and sector leaders cannot articulate needs or evaluate AI solutions, hindering effective adoption of climate‑AI tools.
Evidence
“And this trickles down in many ways because then decision makers who are making decisions about AI at an organizational level, at a policymaker, it’s very hard to pinpoint what’s actually needed if you don’t have that basic literacy.” [46].
Major discussion point
Building capacity and talent pipeline for climate‑AI integration
Topics
Capacity development | Artificial intelligence | Human rights and the ethical dimensions of the information society
Swetha Ravi Kumar
Speech speed
185 words per minute
Speech length
813 words
Speech time
263 seconds
Common technical specifications and standards for interoperability
Explanation
A suite of shared specifications and standards is needed so that all stakeholders can speak a common data language, enabling seamless data exchange across sectors.
Evidence
“So it’s going to be a suite of specifications and standards that allow for all of us to have sort of a common data language.” [61]. “And then it’s for the ecosystem to take and scale with the stakeholders.” [45].
Major discussion point
Data ecosystem challenges and need for granular, interoperable data
Topics
Data governance | Artificial intelligence | Capacity development
AAA framework drives sustained tool adoption
Explanation
The Architecture‑Adoption‑Accelerators (AAA) framework provides a structured pathway—technical specs, adoption incentives, and sandbox accelerators—to embed climate‑AI tools into long‑term workflows.
Evidence
“I’m going to break it down through what we call the AAA framework at the Indie Advertising Stack.” [108]. “I love the AAA framework.” [106].
Major discussion point
Translating data into policy, scaling solutions, and institutional shifts
Topics
Artificial intelligence | Capacity development | Enabling environment for digital development
Simple, language‑local interfaces lower barriers for end‑users
Explanation
Providing tools that operate in local languages and familiar platforms (e.g., WhatsApp) makes climate‑AI services accessible to non‑technical users, fostering broader adoption.
Evidence
“As a consumer, they could talk in their local language to an AI bot in WhatsApp and trade power.” [130]. “…simple WhatsApp based interfaces because they didn’t want to understand all of this complicated AI.” [82].
Major discussion point
Building capacity and talent pipeline for climate‑AI integration
Topics
Capacity development | Artificial intelligence | Closing all digital divides
India Energy Stack builds interoperable digital public infrastructure for power trading
Explanation
The India Energy Stack creates a UPI‑like digital public infrastructure for electricity, enabling seamless, cross‑state power trading and supporting climate‑energy policy implementation.
Evidence
“India Energy Stack Program… similar to UPI for power… can sell electricity from their rooftop power plant to someone in Ladakh.” [143]. “Ministry of Power is bringing in a new national data policy framework for the power sector because we’re talking about critical infrastructure here.” [96].
Major discussion point
Case studies illustrating impact
Topics
Environmental impacts | Data governance | Artificial intelligence
Dr. Priya Donti
Speech speed
188 words per minute
Speech length
711 words
Speech time
226 seconds
Cross‑functional skill gaps hinder effective use of climate‑AI data
Explanation
Effective climate‑AI solutions require coordinated cross‑functional expertise; gaps in such skills limit the ability to translate data into actionable solutions.
Evidence
“Coordination, but I think also being principled about what kind of cross‑functional skills are necessary to actualize and measure solutions…” [70]. “And this trickles down… decision makers… it’s very hard to pinpoint what’s actually needed if you don’t have that basic literacy.” [46].
Major discussion point
Building capacity and talent pipeline for climate‑AI integration
Topics
Capacity development | Artificial intelligence | Data governance
Defining success metrics and procurement models helps institutionalize solutions
Explanation
Clear success metrics and procurement pathways enable organizations to evaluate, adopt, and scale climate‑AI tools in a systematic, accountable manner.
Evidence
“We also kind of defining what success means means having metrics that are kind of stated that are measured.” [113]. “…there’s… a gap where there isn’t enough capacity to upskill internally nor is there actually a good procurement option… enabling this more diverse ecosystem of solutions providers…” [52].
Major discussion point
Building capacity and talent pipeline for climate‑AI integration
Topics
Monitoring and measurement | Capacity development | Artificial intelligence
Open summer school expands AI‑climate talent pool
Explanation
A virtual summer school provides foundational AI and climate knowledge to participants from diverse backgrounds, helping to grow the pipeline of climate‑AI practitioners.
Evidence
“Climate Change AI is running an open registration virtual summer school… provide some of these AI basics as well as climate basics…” [75].
Major discussion point
Building capacity and talent pipeline for climate‑AI integration
Topics
Capacity development | Artificial intelligence | Education
Priyank Hirani
Speech speed
141 words per minute
Speech length
997 words
Speech time
424 seconds
Measure progress with quantitative and qualitative indicators
Explanation
Tracking both quantitative metrics and qualitative feedback is essential to assess the impact of climate‑AI initiatives and guide iterative improvements.
Evidence
“and how do we start measuring these things both quantitatively and qualitatively so that we are able to track the progress.” [124].
Major discussion point
Building capacity and talent pipeline for climate‑AI integration
Topics
Monitoring and measurement | Capacity development
Ensuring coordination, incentives, and capacity for sustained adoption
Explanation
Effective data‑driven climate solutions require coordinated governance, appropriate incentives, and strong institutional capacity across sectors.
Evidence
“So I’m hearing sort of ensuring coordination between departments, ensuring thinking about the data strategy.” [58]. “But amongst these things of standards, interoperability, incentives, what do you think ensures sustained adoption of tools?” [62]. “I think that idea of reducing barriers to access and making any tool user friendly is super critical.” [66].
Major discussion point
Translating data into policy, scaling solutions, and institutional shifts
Topics
Data governance | Capacity development | Enabling environment for digital development
Agreements
Agreement points
Data fragmentation and interoperability challenges
Speakers
– Dr. Cormekki Whitley
– Akhilesh Magal
– Srinivas Krishnaswamy
Arguments
Data fragmentation and lack of interoperability across systems
Unstructured and non-interoperable power sector data with nomenclature inconsistencies
Manual data entry requirements preventing real-time integration
Summary
All three speakers identify significant challenges with data fragmentation, lack of standardization, and interoperability issues across climate and energy data systems, preventing effective integration and real-time access
Topics
Data governance | Information and communication technologies for development | Environmental impacts
Need for standardized data architecture and API access
Speakers
– Akhilesh Magal
– Srinivas Krishnaswamy
– Swetha Ravi Kumar
Arguments
Need for API access and standardized data acquisition methods
Manual data entry requirements preventing real-time integration
Importance of common data language and interoperability specifications
Summary
These speakers agree on the critical need for standardized data architectures, API access, and common data languages to enable efficient data sharing and reduce manual processes
Topics
Data governance | Information and communication technologies for development | Digital Infrastructure and Standardization
Importance of cross-functional skills and capacity building
Speakers
– Dr. Cormekki Whitley
– Dr. Priya Donti
– Priyank Hirani
Arguments
Need for interdisciplinary talent that can translate across climate and AI domains
Need for cross-functional skills and specialized solutions providers for specific sectors
Need for socio-technical skills and cross-functional capacity building
Summary
All three speakers emphasize the critical need for developing cross-functional skills that bridge domain expertise with AI/data capabilities, enabling effective translation between technical and sectoral knowledge
Topics
Capacity development | Artificial intelligence | Information and communication technologies for development
Coordination and multi-stakeholder collaboration requirements
Speakers
– Dr. Cormekki Whitley
– Swetha Ravi Kumar
– Speaker 1
Arguments
Need for interdisciplinary talent that can translate across climate and AI domains
Coordination at scale through whole-of-ecosystem approach using AAA framework
Need for coordination between multiple data-collecting agencies at national and state levels
Summary
These speakers agree that effective climate and AI solutions require extensive coordination across multiple stakeholders, agencies, and sectors, moving beyond siloed approaches to ecosystem-wide collaboration
Topics
The enabling environment for digital development | Information and communication technologies for development | Data governance
Inclusive and equitable transition planning
Speakers
– Dr. Srikant K. Panigrai
– Swetha Ravi Kumar
Arguments
Inclusive planning ensuring nobody is left behind in energy transition
Making AI tools accessible through simple interfaces like WhatsApp for end users
Summary
Both speakers emphasize the importance of ensuring that climate and energy transitions are inclusive, accessible to all stakeholders regardless of their technical capacity, and designed to work for billions while maintaining equity
Topics
Closing all digital divides | Social and economic development | Environmental impacts
Similar viewpoints
Both speakers from the same organization present complementary arguments about heat being a critical macroeconomic variable with severe productivity and health impacts, requiring data-driven localized solutions
Speakers
– Karan Shah
– Professor Neelanjan Sircar
Arguments
Heat as a structural phenomenon requiring macroeconomic consideration beyond meteorological variables
Significant productivity losses and health impacts from heat exposure variations
Topics
Environmental impacts | Social and economic development
Both speakers emphasize the need for clear success definitions and systematic approaches to move beyond pilot projects to sustainable, institutionalized solutions
Speakers
– Dr. Priya Donti
– Priyank Hirani
Arguments
Importance of defining success metrics and solution frameworks from the outset
Moving from pilots to permanent solutions through institutionalization
Topics
The enabling environment for digital development | Monitoring and measurement | Capacity development
Both speakers advocate for comprehensive, integrated approaches to data systems that can unlock significant economic value through unified platforms and holistic perspectives
Speakers
– Srinivas Krishnaswamy
– Akhilesh Magal
Arguments
Integration of climate data with economic indicators for holistic development perspective
Digital public infrastructure design for energy sector similar to UPI for banking
Topics
Information and communication technologies for development | The digital economy | Environmental impacts
Unexpected consensus
AI literacy as fundamental requirement for decision-makers
Speakers
– Dr. Priya Donti
– Priyank Hirani
Arguments
AI literacy requirements at scale among policymakers, NGOs, and industry
Need for socio-technical skills and cross-functional capacity building
Explanation
Despite coming from different backgrounds (MIT academic and capacity building practitioner), both speakers unexpectedly converge on the critical need for basic AI literacy among decision-makers as a prerequisite for effective AI implementation
Topics
Capacity development | Artificial intelligence | Digital literacy
User-centric design for complex technical systems
Speakers
– Swetha Ravi Kumar
– Dr. Priya Donti
Arguments
Making AI tools accessible through simple interfaces like WhatsApp for end users
Need for cross-functional skills and specialized solutions providers for specific sectors
Explanation
An energy infrastructure expert and an AI researcher unexpectedly align on the importance of making complex AI systems accessible through simple, user-friendly interfaces rather than expecting users to understand technical complexity
Topics
Closing all digital divides | Artificial intelligence | Information and communication technologies for development
Scale mismatch in current planning approaches
Speakers
– Karan Shah
– Srinivas Krishnaswamy
Arguments
Scale mismatch between district-level heat action plans and neighborhood-level heat experience
Need for granular data collection and compilation at higher frequency
Explanation
A policy researcher focused on heat impacts and a data dashboard expert unexpectedly converge on identifying scale mismatches as a fundamental problem in current planning approaches, requiring more granular, localized solutions
Topics
Social and economic development | Environmental impacts | Data governance
Overall assessment
Summary
The speakers demonstrate strong consensus on fundamental challenges around data fragmentation, the need for standardized architectures, cross-functional capacity building, multi-stakeholder coordination, and inclusive design. There is remarkable alignment across different sectors and expertise areas on the importance of moving from technical pilots to institutionalized solutions.
Consensus level
High level of consensus with complementary perspectives rather than conflicting viewpoints. The agreement spans technical, governance, and social dimensions, suggesting a mature understanding of the systemic nature of climate and AI challenges. This consensus provides a strong foundation for coordinated action and suggests that the barriers to implementation are well-understood across the ecosystem.
Differences
Different viewpoints
Approach to data standardization and interoperability
Speakers
– Akhilesh Magal
– Swetha Ravi Kumar
Arguments
Building unified and scalable data architecture for India’s power sector
Importance of common data language and interoperability specifications
Summary
Akhilesh focuses on technical data architecture and API solutions for machine readability, while Swetha emphasizes a broader AAA framework that includes specifications and standards but goes beyond pure technical solutions to include adoption pathways and stakeholder engagement
Topics
Data governance | Information and communication technologies for development
Scale and scope of heat action planning
Speakers
– Karan Shah
– Professor Neelanjan Sircar
Arguments
Scale mismatch between district-level heat action plans and neighborhood-level heat experience
Need for localized heat action plans integrating individual and built environment characteristics
Summary
While both agree on the need for more granular planning, Karan emphasizes the institutional scale mismatch problem, while Neelanjan focuses more on the technical integration of individual characteristics and built environment data
Topics
Social and economic development | Environmental impacts | The enabling environment for digital development
Unexpected differences
Role of private versus public adaptation strategies
Speakers
– Karan Shah
– Dr. Srikant K. Panigrai
Arguments
Cooling as private adaptation strategy without adequate public response
Inclusive planning ensuring nobody is left behind in energy transition
Explanation
While both speakers are concerned about equity, Shah identifies cooling as a problematic private adaptation strategy highlighting the lack of public solutions, while Panigrai focuses on ensuring public policy includes everyone in the transition. This represents different views on whether private adaptation is inherently problematic or whether it can be part of an inclusive approach
Topics
Social and economic development | Environmental impacts | Human rights and the ethical dimensions of the information society
Overall assessment
Summary
The speakers show remarkable consensus on identifying core challenges (data fragmentation, need for capacity building, importance of inclusivity) but differ significantly in their proposed solutions and implementation approaches
Disagreement level
Low to moderate disagreement level with high convergence on problem identification but divergent solution pathways. This suggests a mature discussion where stakeholders understand the challenges but bring different expertise and perspectives to solving them. The implications are positive as it indicates multiple complementary approaches rather than fundamental conflicts, suggesting potential for integrated solutions that combine technical, institutional, and social approaches
Partial agreements
Partial agreements
All speakers agree that data fragmentation and lack of interoperability are major challenges, but they propose different solutions: Whitley focuses on building interdisciplinary capacity and partnerships, Krishnaswamy emphasizes the need for digital integration and API access, while Magal advocates for unified data architecture and standardization
Speakers
– Dr. Cormekki Whitley
– Srinivas Krishnaswamy
– Akhilesh Magal
Arguments
Data fragmentation and lack of interoperability across systems
Manual data entry requirements preventing real-time integration
Unstructured and non-interoperable power sector data with nomenclature inconsistencies
Topics
Data governance | Information and communication technologies for development | The enabling environment for digital development
All agree on the critical need for capacity building and skills development, but differ in approach: Donti emphasizes AI literacy and specialized solutions providers, Hirani focuses on socio-technical ‘bilingual’ skills, while Panigrai concentrates on inclusive transition planning for vulnerable workers
Speakers
– Dr. Priya Donti
– Priyank Hirani
– Dr. Srikant K. Panigrai
Arguments
Need for cross-functional skills and specialized solutions providers for specific sectors
Need for socio-technical skills and cross-functional capacity building
Training and capacity building for workers transitioning from fossil to renewable energy
Topics
Capacity development | Artificial intelligence | Social and economic development
Both speakers agree on the importance of inclusivity in digital solutions, but approach it differently: Swetha focuses on technical accessibility through simple interfaces and local language support, while Panigrai emphasizes policy-level inclusive planning that ensures equitable participation in the energy transition
Speakers
– Swetha Ravi Kumar
– Dr. Srikant K. Panigrai
Arguments
Making AI tools accessible through simple interfaces like WhatsApp for end users
Inclusive planning ensuring nobody is left behind in energy transition
Topics
Closing all digital divides | Social and economic development | Environmental impacts
Similar viewpoints
Both speakers from the same organization present complementary arguments about heat being a critical macroeconomic variable with severe productivity and health impacts, requiring data-driven localized solutions
Speakers
– Karan Shah
– Professor Neelanjan Sircar
Arguments
Heat as a structural phenomenon requiring macroeconomic consideration beyond meteorological variables
Significant productivity losses and health impacts from heat exposure variations
Topics
Environmental impacts | Social and economic development
Both speakers emphasize the need for clear success definitions and systematic approaches to move beyond pilot projects to sustainable, institutionalized solutions
Speakers
– Dr. Priya Donti
– Priyank Hirani
Arguments
Importance of defining success metrics and solution frameworks from the outset
Moving from pilots to permanent solutions through institutionalization
Topics
The enabling environment for digital development | Monitoring and measurement | Capacity development
Both speakers advocate for comprehensive, integrated approaches to data systems that can unlock significant economic value through unified platforms and holistic perspectives
Speakers
– Srinivas Krishnaswamy
– Akhilesh Magal
Arguments
Integration of climate data with economic indicators for holistic development perspective
Digital public infrastructure design for energy sector similar to UPI for banking
Topics
Information and communication technologies for development | The digital economy | Environmental impacts
Takeaways
Key takeaways
Moving from pilot innovations to system-level change requires coordination at scale through a whole-of-ecosystem approach that brings all stakeholders to the design board from the beginning
Climate and energy data ecosystems face critical challenges including fragmentation, lack of interoperability, inconsistent nomenclature, and reluctance to share even non-sensitive data
Heat is no longer just a meteorological variable but a macroeconomic one, requiring localized heat action plans that integrate individual characteristics, built environment, and geographic factors
Success requires defining clear metrics and solution frameworks upfront, along with cross-functional skills that bridge domain expertise with AI/data capabilities
Digital public infrastructure for energy (India Energy Stack) should follow the UPI model for banking, enabling simple interfaces for end users while maintaining technical sophistication in the backend
AI literacy at scale is essential among policymakers, NGOs, and industry leaders to enable informed decision-making about AI integration
Just transition principles must ensure inclusive planning where workers transitioning from fossil fuels to renewable energy receive adequate training and livelihood security
Data standardization and API access are crucial for reducing manual intervention and enabling real-time integration across multiple agencies and systems
Resolutions and action items
Climate Change AI will run an open registration virtual summer school focused on AI basics and climate fundamentals to improve cross-sector collaboration
Continue development of the India Energy Stack program with Ministry of Power to create a national data policy framework for the power sector
Expand the unified data architecture approach demonstrated in Goa to other states for power sector data standardization
Develop reference implementation architectures through sandbox environments to demonstrate practical applications for different stakeholders
Create different adoption pathways for stakeholders at varying technical capability levels to ensure inclusive participation
Build more granular data collection and compilation systems at higher frequency to support real-time decision making
Unresolved issues
How to overcome institutional reluctance and sluggishness in sharing non-sensitive data across agencies
Scaling manual data entry processes to automated systems with API integration across all relevant departments
Developing specialized solutions providers for specific sectors rather than generic AI/data providers
Creating sustainable funding and governance models for long-term maintenance of data infrastructure
Establishing clear data access policies and safeguards for critical infrastructure while maintaining openness
Bridging the gap between technical AI capabilities and practical implementation in resource-constrained environments
Measuring and tracking progress on capacity building initiatives quantitatively and qualitatively
Ensuring equitable access to AI tools and benefits across different socioeconomic groups
Suggested compromises
Balance between data openness and security by creating tiered access systems where non-sensitive data is freely available while critical infrastructure data has appropriate safeguards
Combine automated data collection through APIs with continued manual processes during transition periods to maintain data quality
Use simple interfaces like WhatsApp for end users while maintaining sophisticated technical architecture in the background
Create hybrid approaches that leverage both in-house capacity building and external specialized solutions providers
Develop sector-specific standards and specifications rather than universal standards that may not fit all use cases
Implement gradual transition programs for workers moving from fossil fuel to renewable energy sectors with parallel training and employment opportunities
Thought provoking comments
Heat no longer is just a meteorological variable, but is now a significantly important macroeconomic variable.
Speaker
Karan Shah
Reason
This reframes heat from a weather phenomenon to an economic issue, fundamentally shifting how we should approach heat management. It connects climate impacts directly to productivity, competitiveness, and economic growth, making the case for urgent action beyond just environmental concerns.
Impact
This comment established the economic foundation for the entire presentation and shifted the discussion from viewing heat as an environmental challenge to understanding it as a critical economic policy issue. It set up the framework for discussing productivity losses, grid management, and the need for granular heat action plans.
Cooling has become a private adaptation strategy. We still don’t have a public one.
Speaker
Karan Shah
Reason
This insight reveals a critical gap in climate adaptation policy – that individuals are forced to bear the cost and responsibility of heat adaptation while public infrastructure and policy responses remain inadequate. It highlights systemic inequality in climate resilience.
Impact
This observation deepened the conversation about equity and justice in climate adaptation, leading to discussions about grid planning implications and the need for more coordinated public responses. It also connected to later discussions about inclusive solutions and ensuring nobody is left behind.
Without doing this kind of measurement I might be able to look at energy flows over the last two years and guess what the next month of grid load will look like. But it’s going to be very hard to predict three years down the line, five years down the line unless you know who’s using an AC, how much they’re using the AC.
Speaker
Professor Neelanjan Sircar
Reason
This comment articulates the fundamental limitation of current grid planning approaches and makes a compelling case for granular, behavioral data collection. It connects individual behavior patterns to infrastructure planning in a way that demonstrates the necessity of integrated data systems.
Impact
This shifted the discussion from presenting research findings to demonstrating practical applications for infrastructure planning. It bridged the gap between academic research and real-world policy needs, setting up the subsequent presentations about data architecture and energy systems.
Can we do something similar for the Indian power sector where someone from Tamil Nadu can sell electricity from their rooftop power plant to someone in Ladakh, if this is possible I think our work as researchers is really would come to fruition
Speaker
Akhilesh Magal
Reason
This vision statement captures the transformative potential of standardized data architecture by drawing a parallel to UPI’s success in financial services. It makes the technical work tangible and demonstrates how data standardization can enable entirely new economic possibilities.
Impact
This comment elevated the discussion from technical data challenges to envisioning systemic transformation. It provided a concrete, relatable example that helped frame the importance of the technical work being presented and connected to later discussions about coordination at scale.
How do we move from pilot to system level change? How do we design ecosystems that drive adoption, not just innovation? And how do we build the interdisciplinary talent that can translate across climate and AI?
Speaker
Dr. Cormekki Whitley
Reason
These three questions identify the core challenge facing climate-AI initiatives – the gap between proof-of-concept and sustainable implementation. They shift focus from technical capabilities to systemic change requirements, highlighting the need for institutional transformation.
Impact
These questions became the organizing framework for the entire panel discussion that followed. They redirected the conversation from showcasing technical solutions to examining the enabling conditions, governance structures, and capacity building needed for sustained impact.
We need all of them at the design board if we don’t articulate what is in it for me for every stakeholder from early on the question that you asked can we move from pilots to scale would be a recurring question
Speaker
Swetha Ravi Kumar
Reason
This insight addresses a fundamental flaw in many technology initiatives – treating stakeholder engagement as an afterthought rather than a design principle. It emphasizes that sustainable adoption requires understanding and addressing diverse stakeholder motivations from the beginning.
Impact
This comment introduced the critical concept of stakeholder-centered design and influenced the discussion toward practical implementation strategies. It connected to subsequent discussions about inclusivity, coordination mechanisms, and the need for diverse pathways for different types of organizations.
I think often we’re building without doing that. And it leads to things like when we have, let’s say, a pilot innovation, we don’t know where we’re headed. We don’t know what that intermediate success is leading to a final success since we don’t set up stages for actually moving things forward.
Speaker
Dr. Priya Donti
Reason
This observation diagnoses a systemic problem in innovation projects – the lack of clear success metrics and pathways to scale. It challenges the common practice of building solutions without clearly defined objectives or measurement frameworks.
Impact
This comment introduced a more rigorous, systematic approach to solution development and influenced the discussion toward the importance of measurement, evaluation, and strategic planning. It added analytical depth to the conversation about moving from pilots to permanent solutions.
The number of people who could actually define what AI is and what an AI pipeline looks like is extremely small. And this trickles down in many ways because then decision makers who are making decisions about AI at an organizational level, at a policymaker, it’s very hard to pinpoint what’s actually needed if you don’t have that basic literacy.
Speaker
Dr. Priya Donti
Reason
This comment exposes a fundamental contradiction – widespread AI adoption discussions happening without basic AI literacy among decision-makers. It highlights how the lack of foundational knowledge creates barriers to effective implementation and governance.
Impact
This observation brought the discussion full circle to the importance of capacity building and education. It reinforced earlier themes about interdisciplinary skills and provided a concrete call to action for addressing knowledge gaps at the decision-making level.
Overall assessment
These key comments fundamentally shaped the discussion by progressively expanding the scope from technical solutions to systemic transformation challenges. The conversation evolved through several phases: first establishing the economic and social urgency of climate-data integration (Shah’s comments about heat as macroeconomic variable), then demonstrating technical possibilities (Sircar and Magal’s infrastructure examples), and finally examining the institutional and capacity requirements for sustainable implementation (Whitley’s framing questions and the panel responses). The most impactful comments consistently connected technical capabilities to human needs and institutional realities, preventing the discussion from remaining purely technical. They introduced critical concepts like stakeholder-centered design, the need for clear success metrics, and the importance of basic literacy among decision-makers. Together, these insights created a comprehensive framework for understanding why many climate-AI initiatives remain at pilot stage and what systemic changes are needed to achieve scale and sustainability.
Follow-up questions
How do we move from pilot to system level change?
Speaker
Dr. Cormekki Whitley
Explanation
This is a fundamental challenge in scaling climate and AI solutions from experimental phases to widespread implementation that can create systemic impact.
How do we design ecosystems that drive adoption, not just innovation?
Speaker
Dr. Cormekki Whitley
Explanation
Understanding the difference between creating innovative solutions and ensuring they are actually adopted and used by stakeholders is critical for real-world impact.
How do we build the interdisciplinary talent that can translate across climate and AI?
Speaker
Dr. Cormekki Whitley
Explanation
There’s a need for professionals who can work at the intersection of climate science and AI technology, requiring specialized cross-functional skills.
How do we predict grid load management three to five years down the line without detailed individual usage data?
Speaker
Professor Neelanjan Sircar
Explanation
Long-term energy planning requires understanding individual behavior patterns around air conditioning and cooling usage, which current data collection methods don’t adequately capture.
How can we make heat action plans more granularly informed at the neighborhood level rather than state or district level?
Speaker
Karan Shah
Explanation
There’s a scale mismatch between how heat is experienced (neighborhood level) and how heat action plans are currently designed (state/district level).
How do we build API access to government data portals to improve efficiency over web scraping?
Speaker
Akhilesh Magal
Explanation
Current data collection methods are inefficient and require significant manual intervention; API access would enable more automated and reliable data acquisition.
How do we address data silos that contain publicly available information but remain inaccessible?
Speaker
Akhilesh Magal
Explanation
There’s a need to identify and address barriers that prevent access to data that should be publicly available for research and policy purposes.
How do we develop more specialized solutions providers that understand domain-specific needs rather than generic AI providers?
Speaker
Dr. Priya Donti
Explanation
There’s a gap between generic AI solutions and the specialized knowledge needed for specific sectors like power systems, health, or buildings.
How do we provide alternative training and livelihood security for workers transitioning from coal-based to renewable energy jobs?
Speaker
Dr. Srikant K. Panigrai
Explanation
Just transition requires ensuring that workers in fossil fuel industries have pathways to employment in the renewable energy sector.
How do we achieve real-time data integration instead of the current 3-4 day delay in climate and energy dashboards?
Speaker
Srinivas Krishnaswamy
Explanation
Moving from near real-time to real-time data would significantly improve the utility of climate and energy monitoring systems.
How do we overcome reluctance and sluggishness in data sharing, even for non-sensitive data?
Speaker
Srinivas Krishnaswamy
Explanation
Institutional barriers to data sharing are preventing more effective coordination and analysis in the climate and energy sector.
How do we scale AI literacy among policymakers, NGOs, and industry stakeholders?
Speaker
Dr. Priya Donti
Explanation
Decision makers need basic AI literacy to make informed decisions about AI implementation at organizational and policy levels.
Disclaimer: This is not an official session record. DiploAI generates these resources from audiovisual recordings, and they are presented as-is, including potential errors. Due to logistical challenges, such as discrepancies in audio/video or transcripts, names may be misspelled. We strive for accuracy to the best of our ability.
Related event
