Keynote-Martin Schroeter
19 Feb 2026 14:15h - 14:30h
Keynote-Martin Schroeter
Summary
At the AI Summit in India, Martin Schroeter, CEO of Kindrill, urged a shift from AI demos to reliable production systems [5-7]. He said the barrier is not lack of innovation-AI is “brilliant”-but a readiness problem that prevents real-world impact [20-22]. Global research shows over two-thirds of firms invest in AI yet almost half see limited returns, and in India 75 % stall after proof-of-concept [22-24]. Kindrill’s customers seek answers to readiness questions: handling fragmented data, ensuring 24/7 operation, integrating agentic AI in regulated settings, and preparing the workforce [30-41]. Trust, he argued, requires clear guardrails, accountability, transparency and explainability, especially for governments and banks [44-46]. India serves as a proving ground, with the Unified Lending Interface that shortens loan processing from weeks to minutes [50-54]. Kindrill built scalable platforms for banking, telecoms and airports, including an agentic-AI system at Bangalore Airport that enables proactive, self-healing IT operations [56-58]. The firm also supports community skill programmes and is opening a cyber-defence centre in Bangalore to address emerging AI-driven threats [59-60]. He urged moving AI governance into live systems by embedding auditability, logging, explainability and compliance, using “policy as code” for guardrails [65-68]. He noted AI’s impact will be judged not only by productivity gains but by how institutions help people adapt to new automation [71-73]. Building trust, reskilling workers at scale, and ensuring AI aligns with societal values are responsibilities shared by companies and governments [78-81]. Closing the gap between experimentation and industrialisation with infrastructure, security, governance and skilled people is essential for AI to deliver benefits for people, planet and progress [69-70][77].
Keypoints
Major discussion points
– AI readiness, not innovation, is the bottleneck – While AI technology is “brilliant,” most organizations struggle to move beyond proof-of-concept because the supporting infrastructure, data, operations, and people are not yet industrialized for large-scale, reliable deployment [20-22][24-28][30-34].
– Four core readiness questions dominate customer concerns: (1) how to deploy AI across fragmented, multi-cloud and edge data sources; (2) whether AI systems can run 24/7 with resilience to cyber-attacks, outages, data drift and regulatory scrutiny; (3) the suitability of agentic AI for mission-critical, regulated environments; and (4) how to prepare the workforce for new AI-augmented ways of working [30-41].
– India as a strategic proving ground for industrialized AI – The speaker highlights national initiatives (Digital India, India AI Mission) and concrete deployments such as the Unified Lending Interface and agentic AI at Bangalore International Airport, illustrating how AI can be scaled responsibly across public services, finance, healthcare, transport and energy [50-58][60-62].
– Embedding governance, trust and “policy as code” into live AI systems – Trust is built through clear guardrails, auditability, explainability and compliance baked into AI operations, shifting governance from static policy documents to executable code that regulators, boards and citizens can rely on [44-48][65-68].
– Call to action for infrastructure, security, skills and joint responsibility – The speaker urges immediate focus on scalable infrastructure, robust security, workforce reskilling, and collaborative stewardship between companies and governments to close the gap between AI experimentation and industrialization [69-73][78-83].
Overall purpose / goal
The discussion aims to reframe the AI conversation from hype-driven optimism to a pragmatic, “industrialization” mindset. By sharing Kindrill’s experience and research, the speaker seeks to persuade policymakers, business leaders, and technologists that responsible, large-scale AI deployment hinges on readiness-robust infrastructure, governance, reliability, and a prepared workforce-and that coordinated action now will determine AI’s societal impact.
Overall tone and its evolution
– Opening (0:00-5:00) – Formal, appreciative, and optimistic, thanking leaders and emphasizing the opportunity to shape AI responsibly [7-11].
– Middle (5:00-15:00) – Cautiously analytical, highlighting concrete challenges (readiness gaps, stalled projects) and presenting a problem-solving agenda [20-34][30-41].
– Mid-to-late (15:00-25:00) – Inspirational and confidence-building, using India’s initiatives and success stories to illustrate feasible pathways [50-58][60-62].
– Closing (25:00-end) – Urgent, rallying, and forward-looking, issuing a clear call to action and stressing shared responsibility, while maintaining a hopeful note about AI’s transformative potential when industrialized responsibly [69-83][84].
The tone shifts from respectful acknowledgment to critical assessment, then to hopeful illustration, and finally to a decisive, motivational appeal.
Major discussion points
– Readiness, not innovation, is the bottleneck – AI technology itself is “brilliant,” but most projects stall because the supporting infrastructure, data, operations, and people are not yet industrialized for large-scale, reliable deployment [20-22][24-28][30-34].
– Four core readiness questions dominate customers: (1) deploying AI across fragmented, multi-cloud and edge data; (2) ensuring 24 × 7 reliability, resilience to cyber-attacks, data drift and regulatory scrutiny; (3) assessing the suitability of agentic AI for mission-critical, regulated environments; and (4) preparing the workforce for AI-augmented ways of working [30-41].
– India as a strategic proving ground for industrialized AI – National initiatives (Digital India, India AI Mission) and concrete deployments-such as the Unified Lending Interface and agentic AI at Bangalore International Airport-show how AI can be scaled responsibly across public services, finance, healthcare, transport and energy [50-58][60-62].
– Embedding governance and trust into live AI systems – Trust is built through clear guardrails, auditability, explainability and compliance baked directly into AI operations, moving governance from static policy documents to “policy as code” that regulators, boards and citizens can rely on [44-48][65-68].
– Call to action: infrastructure, security, skills and shared responsibility – Immediate focus is needed on scalable infrastructure, robust security, workforce reskilling, and collaborative stewardship between companies and governments to close the gap between AI experimentation and industrialization [69-73][78-83].
Overall purpose / goal
The speaker’s goal is to shift the AI conversation from hype-driven optimism to a pragmatic, “industrialization” mindset. By sharing Kindrill’s research and real-world examples, he urges policymakers, business leaders, and technologists to prioritize readiness-robust infrastructure, governance, reliability, and a prepared workforce-so that AI can deliver real-world impact at national and enterprise scale.
Overall tone and its evolution
– Opening (0:00-5:00) – Formal, appreciative, and optimistic, thanking leaders and framing the summit as an opportunity to shape AI responsibly [7-11].
– Middle (5:00-15:00) – Cautiously analytical, highlighting concrete challenges (readiness gaps, stalled projects) and laying out a problem-solving agenda [20-34][30-41].
– Mid-to-late (15:00-25:00) – Inspirational and confidence-building, using India’s initiatives and success stories to illustrate feasible pathways [50-58][60-62].
– Closing (25:00-end) – Urgent, rallying, and forward-looking, issuing a decisive call to action and emphasizing shared responsibility while maintaining a hopeful note about AI’s transformative potential when industrialized responsibly [69-83][84].
The tone moves from respectful acknowledgment to critical assessment, then to hopeful illustration, and finally to a decisive, motivational appeal.
Speakers
– Martin Schroeter – Role/Title: Chairman and CEO, Kyndryl (referred to as Kindrill in the transcript) – Area of Expertise: IT infrastructure services, AI operationalization, enterprise technology [S2].
– Speaker 1 – Role/Title: Event moderator/host (introducing the keynote speaker) – Area of Expertise: (not specified) [S4].
Additional speakers:
– (none)
The session opened with Speaker 1 introducing Martin Schroeter as chairman and CEO of Kindrill, the world’s largest IT-infrastructure services company spun out of IBM, and noting that his perspective would temper the summit-stage optimism surrounding AI [1-4]. Schroeter then thanked Prime Minister Narendra Modi for convening the gathering of ministers, policymakers, CEOs and the global audience, and stressed the extraordinary opportunity to shape a new era of AI that is responsible for people, industry and communities [5-10]. He positioned Kindrill’s engineers, consultants and mission-critical support teams as the collective knowledge base behind the discussion [11-12].
Schroeder quickly reframed the conversation from hype to pragmatic readiness, arguing that the barrier to real-world AI impact is not a lack of innovation-AI is “brilliant”-but a readiness problem that prevents industrialisation [20-22]. Global studies show that while more than two-thirds of organisations are heavily invested in AI, almost half struggle to achieve meaningful returns, and in India 75 % of projects stall after the proof-of-concept stage [22-24]. According to Kindrill’s experience, the leading cause of stall is not the technology itself but the absence of an industrialised ecosystem of infrastructure, data, operations and people [25-28].
He identified four core readiness questions that dominate customers’ concerns: first, how to deploy AI when data is fragmented across multiple clouds, core systems of record and edge environments, especially where business processes were never designed for AI and regulatory regimes differ by sector and geography [30-32]; second, how to ensure AI systems can run 24 × 7 without failure, withstand cyber-attacks, outages, data drift and regulatory scrutiny, and earn user trust [33-35]; third, whether organisations are truly ready to use agentic AI in mission-critical, regulated settings and how such agents can be integrated with existing stacks [36-39]; fourth, how to prepare the workforce for AI-augmented ways of working, given that nine in ten leaders expect AI to reshape work yet fewer than one in three feel their employees are ready [40-42].
Trust is the linchpin linking these challenges; leaders can only rely on AI when it operates within clear, accountable, transparent and explainable guardrails-requirements especially vital for governments, banks and other regulated industries [44-46]. He described these as “core readiness challenges” that cause many AI initiatives to stall, emphasizing that innovation must become reliable, predictable and secure in day-to-day operations [47-49]. Embedding governance directly into live AI systems-through auditability, logging, explainability and compliance-transforms policy from static documents into executable code, a “policy as code” approach that provides concrete guardrails for agentic AI and builds confidence among regulators, boards and citizens [65-68].
India was presented as a strategic proving ground for industrialising AI at massive scale. He emphasized that in India, “scale means something different… failure is not an option” [15-17]. Under Prime Minister Modi’s leadership, the country has elevated AI to a national priority, creating policy, digital and talent foundations such as Digital India and the India AI Mission that support large-scale, inclusive innovation [51-55]. Concrete deployments illustrate this potential: the Unified Lending Interface now reduces loan-approval times from weeks to minutes while expanding credit access, and at Bangalore International Airport Kindrill has applied agentic AI to shift IT operations from reactive to proactive, enabling self-healing capabilities that improve predictability and trust [53-58].
Beyond these pilots, Kindrill is deepening its commitment to India’s AI ecosystem. The company is opening a new cyber-defence operations centre in Bangalore to detect and contain AI-driven threats at the network edge before they cause disruption [60]. Simultaneously, it is expanding community partnerships that build digital and cybersecurity skills, recognising that safe, responsible AI adoption depends as much on people as on technology [59-60]. These initiatives reflect a broader strategy to support scalable platforms for banking, citizen services, telecoms and airports, handling millions of daily users and transactions [56-58].
Schroeter concluded with a clear call to action: stakeholders must focus immediately on the fundamentals-scalable infrastructure, trustworthy security and a skilled workforce-to operationalise AI responsibly [69-73]. He warned that AI’s true impact will be judged not only by productivity gains but by how institutions help societies adapt to the next phase of industrial automation, and that the transition from invention to impact requires joint investment from both companies and governments; only when AI is industrialised safely, reliably and at scale will it strengthen the institutions on which societies depend, rather than merely optimise them [78-83]. He closed by thanking the audience and reaffirming that the future of AI will be decided by the choices and investments made today [84].
Ladies and gentlemen, I would now like to welcome Mr. Martin Schroeter, who is the chairman and CEO, Kindrill. As the leader of the world’s largest IT infrastructure services company spun out of IBM, Mr. Martin Schroeter manages the technology backbone of thousands of enterprises across the globe. His view of what it takes to actually run AI in production environments offers a necessary corrective to summit stage optimism. Ladies and gentlemen, please join me in welcoming the chairman and CEO of Kindrill, Mr. Martin Schroeter.
Thank you. Thank you. Thank you very much. Good afternoon, everybody. First, I want to thank the Honorable Prime Minister of India, Sri Narendra Modi, for convening this distinguished group of ministers, policymakers, global leaders, fellow CEOs, and of course, everybody watching on the live stream. And I want to thank all of you for your support and for your support for the initiative that we are carrying out in this country. And I want to thank all of you for your support and for your support for the initiative that we are carrying out in this country. It is an extraordinary opportunity for us to be here with you as we all focus on how to usher in this new era of AI responsibly for people, for industry, and for our communities.
Today, I’m proud to represent the collective knowledge and experience of Kindrel’s engineers, technical practitioners, problem -solving consultants, the people who support the mission -critical systems that the world depends on every day. As the largest IT infrastructure services provider, the question that we continuously come back to at Kindrel, and one that I suspect many of the policymakers and the business leaders and the technologists and the citizens here among us have, is how do we actually make AI work in the real world for real -world impact? Not a demo, not a pilot or an experiment. And not in theory, but in day -to -day operations under real constraints with people working alongside AI agents at national and enterprise scale.
Scale means something here in India that’s different than anywhere else, where failure of these systems is just not an option. Because when AI moves, when it moves from labs into the systems that power economies, the hospitals and the banks and the transportation networks and the energy grids and the governments, getting it wrong, and these are the systems we run every day, getting it wrong is not just an inconvenience, it actually impacts lives. And these systems sit at the heart of what this summit represents, the people, the planet, and the progress that we’re all working on. Progress in all three depends on the ability to operationalize AI reliably and, again, at scale. So today I’ll share a bit about what we’re learning, working with our global customer base and our partners to close the gap between investments, intelligence and reality, and where AI either becomes part of how we work and how work actually gets done.
or never makes it out of the experimentation phase. And what we’re seeing is not an innovation problem. The innovation is real, but it’s a readiness problem. We’ve conducted global studies with business and IT leaders countless times, and our research shows that while more than two -thirds of global organizations are already heavily invested in AI, almost half still struggle to see meaningful returns. And in India, in India alone, 75 % said their innovation efforts stall after the proof -of -concept stage. So based on our research and our experience with our customers, both in regulated and unregulated industries, the reason, the leading indicator for why projects stall is not because of the technology isn’t smart. It’s brilliant. It’s brilliant.
It’s because we haven’t industrialized it yet. AI today is not industrialized. The infrastructure, the data, the operations, and the people simply aren’t ready to support AI adoption and deployment at scale. So our customers really want greater clarity and greater support on four critical questions. First, on operational conduct, they want to know how to deploy AI when data is fragmented across clouds, across their core systems of record, and at the edge of the environments in which they operate. When business processes were never designed for AI, and when regulations differ by sector and by geography, and when trust, security, and resilience are imperative to how it works. Second, and more systemically, they’re asking, can this system really run 24 by 7 without failure?
Can it withstand cyber attacks and outages and data drift and regulatory scrutiny? And can the people trust it when it matters most? And can it? Can they trust the decisions it’s going to make? Those are the systems we run every day. Third, they’re asking about agentic AI. Whether they’re truly ready to use it in mission -critical environments, are they able to meet the regulatory requirements that come with those environments, and are they able to integrate with existing systems? And fourth, they’re asking about their workforce. How to prepare people for new ways of working with AI. Nine in ten leaders expect AI to fundamentally reshape work, yet fewer than one in three believe their workforce is ready.
Or that they’re equipped to help their teams get there. All of this ladders up to trust. Can leaders trust these AI systems and the insights they provide? And that trust is built when AI operates within clear guardrails where actions are accountable and transparent and explainable, which is essential for organizations in every industry, and especially in government, in banking, and other regulated environments. These are the core readiness questions. And the core readiness challenges that we see every day. And they’re at the heart of why so many AI initiatives stall. They remind us that innovation must operate reliably, predictably, and securely, day after day, in the real world. So I’m thrilled that this year’s AI Summit is India because India is one of the world’s most important proving grounds for industrializing AI at extraordinary scale.
Under the leadership of Prime Minister Modi, India has recognized AI as a strategic national priority, building policy and digital and talent foundations needed to support innovation, and again, at scale. Through initiatives like Digital India and the India AI Mission, and investments in digital public infrastructure, India has positioned itself not just as an adopter of AI, but as a global contributor to how AI can be deployed responsibly and inclusively. AI -powered platforms like the Unified Lending Interface are expanding access to credit at scale, reducing loan times from weeks to minutes, and while improving transparency and inclusion. India’s digital experience offers an important lesson for the world when technology must operate at a national scale across public services and financial systems, healthcare, transportation, and energy.
Reliability, governance, and human integration are not features, they are prerequisites. Kindle is very proud to be a partner to many of India’s leading companies and government agencies. Our local engineering teams have built scalable platforms for banking, for citizen services, for telecoms, and for airports to handle the millions of users and transactions every day. At Bangalore International Airport, we’ve applied agentic AI to shift IT operations from a reactive response to a proactive resilience, supporting self -healing capabilities that improve operational predictability and strengthen trust in the airport’s digitalization. Through our community partnerships in India, we’re helping build digital and cybersecurity skills because safe, responsible AI adoption depends on people being ready. not just technology. And because sophisticated adversaries are already using AI to move at machine speed tomorrow, tomorrow we’re opening a new cyber defense operations center in Bangalore so we can detect and contain threats that already start at the edge of the network before they become disruptions.
So we are deeply committed to helping India and our partners around the world implement AI at the scale to drive people, planet, and progress outcomes. In every part of the globe, conversation about agentic must now shift from intelligence to industrialization, from what AI can do to how it’s orchestrated and how it’s governed and secured and integrated, and how it’s sustained with agents and humans partnering to drive business impact. This is a transition every major technology invention has gone through. Invention comes first, but impact only comes when society’s learned how to industrialize it safely, reliably, and at scale. A critical part of this industrialization is operationalizing the governance of AI. That means moving governance out of policy documents and into live systems, embedding auditability, logging, explainability, and compliance directly into how AI operates.
We’re seeing how our approaches, like policy as code, can establish clear guardrails for agentic AI to drive trust and compliance, giving regulators, boards, and the citizens alike the confidence in these systems are controlled, accountable, and safe. So what do we do next? Excuse me. We get ready by focusing on the fundamentals, infrastructure that can scale, security that earns trust, and people with the skills to operate. We operate AI responsibly. This readiness perspective is particularly important for policymakers. Excuse me. Because the impact of AI cannot be measured only by productivity gains or economic growth. as important as those are to drive the future, it will also be measured by how institutions help people adapt in the next phase of industrial automation and how work evolves.
Excuse me. AI can absolutely change the world. It can change work, it can change skills, it can change mindsets, and it can change operating models. But it will only change, oh, thank you very much, it will only change the world when it is embedded responsibly and reliably into the systems that society depends on every day. The future of AI will not be decided in the research labs or the boardrooms. It will be decided by the choices and the investments we make now, by how we close the gap between experimentation and industrialization. Excuse me. The work ahead is hard, because this is not just a technology shift, it’s a human shift. We have to build trust in AI, we have to reskill our workforces at scale, and we have to ensure these systems are worthy of the societies that depend on them.
The responsibility belongs to the companies and the governments alike. And it is a responsibility worth embracing, because when AI is industrialized responsibly, it doesn’t just optimize. It strengthens the institutions people rely on every day. And that is how AI truly changes the world. Thank you very much.
“Martin Schroeter is chairman and CEO of Kindrill, the world’s largest IT‑infrastructure services company spun out of IBM.”
The knowledge base identifies Martin Schroeter as chairman and CEO of Kyndryl, described as the largest IT infrastructure services provider, confirming his role and the company’s scale [S7].
“Kindrill’s engineers, consultants and mission‑critical support teams constitute the collective knowledge base behind the discussion.”
The source states that Kindrill’s engineers, technical practitioners, consultants and mission-critical support staff represent the collective knowledge and experience for the event [S2].
“In India, 75 % of AI projects stall after the proof‑of‑concept stage.”
The knowledge base reports that almost 80 % of AI pilots fail to reach production, without specifying India, indicating a different percentage and broader scope [S8].
“The leading cause of AI initiative stalls is the absence of an industrialised ecosystem of infrastructure, data, operations and people.”
The source adds that data silos, lack of governance and poor data quality are primary reasons pilots stall, providing more detail on the ecosystem gaps [S8].
“AI systems must be able to run 24 × 7 without failure, withstand cyber‑attacks, outages, data drift and regulatory scrutiny, and earn user trust.”
The transcript excerpt explicitly asks whether AI can withstand cyber attacks, outages, data drift and regulatory scrutiny, and whether people can trust it [S1].
“Organizations need to assess readiness for agentic AI in mission‑critical, regulated settings and how such agents integrate with existing stacks.”
The source notes that a key question from leaders is about agentic AI and whether organizations are truly ready for it [S1].
“Trust is essential; AI must operate within clear, accountable, transparent and explainable guardrails, especially for governments, banks and other regulated industries.”
Multiple sources highlight trust infrastructure as critical, emphasizing transparency, explainability, accountability and security for regulated sectors [S70] and [S77].
“Embedding governance directly into live AI systems—auditability, logging, explainability and compliance—creates a ‘policy as code’ approach that provides concrete guardrails for agentic AI.”
While the source does not mention ‘policy as code’, it does describe four guardrails (fairness, accountability, privacy, security) that are embedded in AI deployments, offering related contextual detail [S91].
The discussion shows strong convergence between the moderator’s framing and Martin Schroeter’s detailed briefing. Both agree that AI’s promise must be grounded in practical readiness – including industrial‑scale infrastructure, trustworthy governance, continuous operation, and skilled people – and that India serves as a strategic proving ground for these efforts. Additional internal consistency in Schroeter’s arguments reinforces a unified narrative around responsible AI industrialization.
High consensus on the need for responsible, industrial‑scale AI deployment and the role of policy, trust, and workforce development. This consensus suggests that future initiatives are likely to prioritize readiness, governance frameworks, and joint public‑private investment rather than purely hype‑driven pilots.
The transcript contains only an introductory segment by Speaker 1 and a single substantive presentation by Martin Schroeter. No other speakers offer contrasting viewpoints, so there are no identifiable points of disagreement or partial agreement within the provided material.
None – the discussion is essentially a monologue presenting a cohesive perspective on AI industrialization, readiness, and joint responsibility. Consequently, the implications for the topic are that the session reinforces a unified industry‑government narrative rather than exposing contested positions.
Martin Schroeter’s remarks transformed a typical summit keynote from a celebratory showcase of AI potential into a grounded, systems‑level critique of readiness. By repeatedly shifting focus—from the novelty of AI, to the unique scale and risk in India, to four concrete readiness questions, and finally to concrete governance mechanisms like ‘policy as code’—he created multiple turning points that redirected the audience’s attention toward operational reliability, trust, and human factors. These thought‑provoking comments not only introduced new ideas but also challenged the prevailing optimism, prompting listeners to reconsider the prerequisites for AI impact and to view industrialization, governance, and workforce preparation as the decisive battlegrounds for responsible AI deployment.
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
