For most technology transitions over the last three decades, the first kind of leader was right. Cloud migration rewarded patience. Mobile commerce gave enterprises years to watch consumer behaviour before demanding a response. The observation window was real, and the enterprises that used it wisely built more durable capabilities than the ones that moved on instinct alone.
Q1. You’ve framed Agentic AI as a question CEOs should have already answered. What makes this different from previous technology cycles, where companies had more time to observe before acting?
Previous technology shifts gave enterprises a runway to watch and wait. Cloud migration took a decade to mature from early adopter to mainstream. Agentic AI does not offer that luxury. Gartner projects that 40 percent of enterprise applications will carry task-specific AI agents by the end of this year, up from less than 5 percent in 2025. That is not a gradual curve, it is a compressed window.
Q2. When you say CEOs need to “answer the question,” what exactly is the question? Is it whether to adopt Agentic AI, or something more specific?
The real question is whether the organisation has decided what Agentic AI is actually meant to do inside their specific business, which processes it will own end to end, what level of autonomy it is given, and who is accountable when it acts. Most CEOs have answered the adoption question. Very few have answered the architecture question.
Q3. Many CEOs believe they have already answered this question because their organisation is running pilots. Why is that often a false signal?
Because a pilot answers a technical question, not a strategic one. It tells you nothing about whether your data is ready, whether your legacy systems can support autonomous execution, or whether your organisation has the governance structure to let an AI agent make a decision without a human checking it first. A pilot proves possibility. It does not prove readiness.
Q4. Based on what you have seen across enterprise engagements, what is the single clearest signal that a company is actually ready for Agentic AI, versus one that is not?
It is whether the company can tell me, with precision, what business outcome the agent is supposed to change and how they will measure it before the project starts. The technology readiness is rarely the gap. The clarity of intent is.
Q5. Indian enterprises have moved fast on Generative AI adoption. Why hasn’t that same speed translated into Agentic AI readiness?
Generative AI is forgiving. If a chatbot gives an imperfect answer, a human catches it before anything happens. Agentic AI removes that safety net, because the agent is meant to act, not just suggest. Indian enterprises built confidence quickly with GenAI because the risk of getting it wrong was low. Agentic AI demands the opposite, you need to be right before you let it run, not after.
Q6. You have spoken about enterprises grafting AI onto legacy processes rather than rebuilding around it. Can you describe what that failure looks like in practice?
A customer asks an AI agent to resolve a complaint, the agent gathers the information beautifully, and then hands it off to the exact same five-step manual approval chain that existed before the AI arrived. Nothing about the underlying workflow changed. The enterprise spent money on AI and kept its old bottleneck. That is not transformation, it is decoration.
Q7. If a CEO came to you today and said, “I am ready to answer this question properly,” what would you tell them to do first?
Pick one process, not ten, and define exactly what success looks like before any technology conversation happens. Get your data architecture assessed honestly, most leaders overestimate how ready their data actually is. Decide who in the organisation owns the outcome, not just the technology rollout.
Q8. What role should the CFO play in this decision, given that ROI measurement has been such a persistent challenge industry-wide?
The CFO has to be in the room from day one, not brought in at the review stage to evaluate what already happened. A CFO who insists on a measurable outcome and a baseline before approving budget changes the entire trajectory of the project.
Q9. You have built INT.’s AI and digital practice around this exact transition, spanning conversational AI, enterprise knowledge intelligence, and workforce learning. What did building this portfolio teach you about what enterprises actually need versus what they think they need?
This pattern shows up with OneSpace, which acts as a company’s brain – and an operating system for AI transformation. The real problem is that the information lives in ten disconnected systems, ERP, CRM, HRMS, documents, email, with no shared layer underneath. OneSpace acts as the second brain of the organisation, connecting all of it so the right answer reaches the right person instantly, with the source cited.
What enterprises think they need is the newest AI capability. What they actually need is integration depth, the unglamorous work of making AI function reliably inside a complex, regulated environment.
Q10. Eighteen months from now, what will separate the Indian enterprises that made the right call on Agentic AI from those that did not?
It will not be who deployed the most agents. It will be who can point to a specific business outcome and say, with data, that it changed because of this technology. Gartner itself expects more than 40 percent of agentic AI projects to be cancelled by 2027 due to unclear value or weak governance. Eighteen months from now, the winners will be the companies that treated this as an architecture decision in 2026, not the ones that treated it as a procurement decision.
The enterprises that will look back on 2026 as the year they got this right will not be the ones that approved the largest AI budgets or launched the most pilots. They will be the ones whose CEOs asked the harder question: not whether we are using Agentic AI, but whether we have built the conditions under which Agentic AI can actually work.
%20(1).jpg)
.jpg.jpeg)
