In the fast-paced arena of enterprise transformation, few leaders carry the credibility and experience of Anil Chintapalli. Over a career spanning more than three decades, Anil has navigated the rare crossroads of global P&L leadership, technology investment, and hands-on operational management. Having overseen multiple public listings and over 20 acquisitions globally unlocking massive returns for his shareholders over the past three decades —including the landmark 2025 $3.3 billion cash acquisition of WNS Holdings by Capgemini—he has earned a reputation for driving substantial return on invested capital through disciplined, large-scale change initiatives.
As Managing Partner at Human Capital Development and Senior Advisor to McKinsey & Company, he is actively shaping how Fortune 500 companies approach artificial intelligence.
In this exclusive interview, we explore why urgency alone cannot sustain transformation, the inner workings of his “Agentic Workforce Operating System,” and why the next frontier of AI is as much about leadership as technology.
Q: Anil, you’ve guided transformations for over 50 Fortune 500 companies and led multi-billion-dollar exits and public listings unlocking substantive shareholder value. Many organizations treat AI as a technical add-on, but you argue it demands a rethink of the operating model. Why do companies struggle to scale AI beyond pilot projects?
Anil Chintapalli: Most enterprises fall into “point-solution thinking.” They implement AI for a narrow task but fail to redesign the workflow around it. Transformation is not about adding a new tool—it’s about achieving outcomes like faster decision-making, improved resilience, or greater efficiency. If AI isn’t embedded into the core operating model, it just adds complexity. The single most obstacle is fragmentation—and I mean that in a very specific, structural sense. Most large enterprises do not have an AI problem; they have an orchestration problem.
Individual departments deploy AI in isolation: marketing builds a recommendation engine, finance builds a fraud detection model, supply chain builds a demand forecasting tool. Each of these may be technically impressive in its own domain. But collectively, they create what I call ‘intelligence silos’—pockets of machine reasoning that cannot communicate, cannot share context, and cannot compound their insights across the enterprise.
This is the organizational equivalent of having brilliant specialists who never speak to one another. The recommendation engine does not know what the fraud model has learned about customer behavior. The demand forecast does not incorporate the signals the marketing team’s AI has detected about shifting consumer sentiment. The result is an enterprise that is locally intelligent but globally incoherent.
Scaling AI requires three foundational elements that most organizations underinvest in. First, you need in-house centers of excellence that own the AI capability horizontally across the organization—not as a shared service, but as an enterprise-grade competency with authority and accountability. Second, you need robust data governance: unified taxonomies, clear ownership models, rigorous quality standards, and interoperable data architectures that allow AI to reason across previously siloed datasets. Third, you need orchestrated workflows—designed end-to-end processes that allow intelligent agents to hand off context, share inferences, and compound their reasoning across functional boundaries.
Without this foundation, AI remains a collection of interesting experiments. With it, AI becomes an enterprise nervous system—a connective tissue that transforms how the entire organization perceives, reasons, and acts. The difference between the two is the difference between a technology cost and a valuation multiplier.
Q: You’ve introduced the “Agentic Workforce Operating System.” How is the system different from traditional automation, and how does it reshape labor structures?
Anil Chintapalli: Traditional automation follows fixed scripts. Agentic AI, in contrast, acts like a digital teammate capable of reasoning through complex, multi-step tasks. The Agentic Workforce Operating System (AWOS) deploys AI squads alongside human teams, directly linking technology to measurable business outcomes. This mixed system lessens the need for expensive consulting, improves workforce efficiency, and changes how we measure success from just using AI to achieving real business results—like speeding up cash-to-order times or lowering revenue risks.
This represents a paradigm shift in how we conceptualize the relationship between human capital and machine intelligence. For the last four decades, the dominant model has been ‘Human and Tool’: a person uses software to accomplish a task. The tool is passive; the human provides all judgment, context, and decision-making authority. AWOS transitions the enterprise to a fundamentally different model.
In my architecture, AI agents are not passive tools awaiting instruction. They are autonomous actors operating within clearly defined protocols, guardrails, and governance frameworks. They can execute multi-step workflows, make bounded decisions, escalate exceptions, and learn from outcomes—all without requiring human intervention at every step. The human role shifts from task execution to strategic orchestration: defining objectives, setting ethical boundaries, interpreting ambiguous situations, and focusing on the high-judgment, high-empathy work that machines cannot replicate.
The productivity implications are profound, but they require a new measurement framework. Traditional productivity metrics—hours logged, tasks completed, output per FTE—are artifacts of the ‘Human and Tool’ era. In the AWOS model, productivity is measured by orchestration effectiveness: how well does the human-agent ecosystem convert inputs into tangible business outcomes? What is the decision velocity? What is the error rate of autonomous agent actions? How quickly can the system adapt when business conditions change? These are the metrics that matter in an agentic enterprise, and they bear almost no resemblance to the productivity frameworks of the previous generation.
Q: With your extensive M&A experience, how does your “investor-operator” perspective inform the way you assess a company’s AI strategy?
Anil Chintapalli: The lens combines rigor and practicality. Investors focus on defensibility and ROI, while operators consider execution risk. I look for leaders with real skin in the game—equity or aligned incentives—because without that, cultural transformation stalls. A strong AI strategy needs a growth playbook anchored in metrics like customer lifetime value. If AI fails to impact core business metrics, it becomes a cost rather than a strategic asset.
Q: Culture is often cited as critical in transformations. How do you mitigate resistance when introducing AI?
Anil Chintapalli: Urgency might mobilize a workforce, but trust and clarity sustain it. Employees pay attention to leadership signals—is collaboration and transparency rewarded, or are silos protected? I invest heavily in upskilling teams, showing that AI handles repetitive tasks and frees humans for higher-value work. When employees see AI enhancing their roles rather than replacing them, resistance diminishes. Transformation, in the end, is a leadership challenge more than a technological one.
Q: You authored a blueprint for SAP adoption at scale and are now co-writing a book on enterprise AI. What connects these two eras of technology?
Anil Chintapalli: The constant is operational integration. Whether it’s SAP or Agentic AI, technology alone doesn’t create value—disciplined adoption does. You need a blueprint that aligns the technology foundation with the operating model. My advice: prioritize progress over perfect information. Don’t chase the latest “model of the month.” Build the system around it. That is how lasting enterprise value is achieved.
Q: What role does data governance play in unlocking AI’s full potential, and why do most enterprises get it wrong?
Anil Chintapalli: Data governance is the single most under-appreciated enabler of enterprise AI, and it is under-appreciated precisely because it is unglamorous. No one wins innovation awards for building a data taxonomy. No one gives keynote speeches about metadata management. But without rigorous data governance, every AI deployment is built on a foundation of noise, inconsistency, and unreliability.
Most enterprises approach data governance as a compliance exercise—something mandated by regulators or auditors. That framing is catastrophically insufficient in the AI era. Data governance is not about compliance; it is about enabling machine reasoning at enterprise scale. When AI agents attempt to reason across datasets that lack consistent definitions, standardized formats, and verified quality, the outputs are unreliable at best and dangerous at worst. The adage ‘garbage in, garbage out’ has never been more consequential than in an environment where autonomous agents are making decisions based on the data they are fed.
The enterprises that get this right treat data governance as a first-class strategic function, not an IT housekeeping task. They appoint senior leaders with genuine authority over data standards. They invest in data engineering infrastructure that rivals their application development investment. They establish clear ownership models so that every dataset has an accountable steward. And critically, they build feedback loops so that AI model performance continuously informs and improves data quality. This creates a virtuous cycle: better data produces better AI outputs, which reveal data quality gaps, which drive further governance improvements. That cycle is the engine of compounding intelligence.
Q: How do you assess the ROI of AI investments, and what metrics do you consider most meaningful?
Anil Chintapalli: Traditional return-on-investment frameworks are necessary but insufficient for evaluating AI. The challenge is that AI generates value across multiple dimensions simultaneously, and many of those dimensions are poorly captured by conventional financial metrics. If you evaluate AI solely on cost reduction, you will miss the revenue acceleration. If you evaluate it solely on productivity gains, you will miss the risk mitigation. The measurement framework must be multidimensional.
I advocate for a three-tier evaluation architecture. The first tier is operational impact: measurable improvements in throughput, error rates, cycle times, and process efficiency. These are the most immediately quantifiable and typically provide the business case for initial investment. The second tier is strategic impact: improvements in decision quality, market responsiveness, customer lifetime value, and competitive positioning. These take longer to materialize but are often more valuable. The third tier is optionality value: the degree to which an AI investment creates future capabilities that do not yet have a defined use case but expand the enterprise’s strategic action space.
Most organizations stop at the first tier, which is why they consistently undervalue AI and underinvest in the foundational capabilities—data governance, talent development, cultural alignment—that enable the second and third tiers. Sophisticated investors understand that the highest-value AI investments are those that expand the enterprise’s option set, even if the immediate P&L impact is modest. This is where the investor-operator lens becomes indispensable: it forces a conversation about both near-term execution and long-term strategic positioning.
Q: What is your perspective on the risks of AI, and how should enterprise leaders manage them responsibly?
Anil Chintapalli: AI risk is real, material, and deserving of serious executive attention. But the discourse around AI risk is often unproductive because it conflates two fundamentally different categories: technical risk and deployment risk. Technical risk—model hallucinations, bias in training data, adversarial vulnerabilities—is important and well-studied. But in my experience, the risks that actually derail enterprise AI programs are deployment risks: misaligned incentives, inadequate governance, premature autonomy, and the absence of meaningful human oversight at critical decision points.
The most dangerous AI deployment is one that operates without clear accountability. When an AI agent makes a consequential decision—approving a loan, flagging a transaction as fraudulent, recommending a medical intervention—there must be an unambiguous chain of accountability that connects the machine output to a human decision-maker who bears responsibility. This is not about slowing AI down; it is about ensuring that the speed of machine reasoning does not outpace the organization’s capacity for responsible governance.
The framework I recommend is ‘graduated autonomy.’ AI agents begin with narrow, well-defined tasks where the consequences of error are limited and reversible. As the organization builds confidence in the agent’s reliability, the scope of autonomy expands incrementally. At every stage, there are clearly defined escalation protocols, human review checkpoints, and audit trails. This approach balances the efficiency gains of AI autonomy with the governance requirements of responsible enterprise operation. It is not glamorous, but it is the only approach that scales without creating catastrophic tail risk.
The Roadmap for Lasting Impact
Anil Chintapalli’s insights underscore a timeless lesson: while the tools of transformation evolve—from ERPs to agentic AI—the fundamentals of leadership and execution endure. His “Investor-Operator” approach reminds us that technology’s impact is only as strong as the operational and cultural scaffolding supporting it.
