Fix The Foundation Then Add Intelligence

India is ahead of the world on enterprise AI adoption. Deloitte's State of AI in the Enterprise 2026 report indicates that about 40% of Indian enterprises are operationalising AI, against a global average of 28%. SAP's Value of AI Report 2025 expects Indian AI investment to rise by 33% over the next two years, with the average Indian firm spending around US$31 million on AI in 2025—above the global benchmark. SAP also found that 93% of Indian businesses expect a positive return within three years.That is the headline. The subtext is more sobering.MIT, in its GenAI Divide: State of AI in Business 2025 study, found that 95% of generative AI pilots globally never make it into production. BCG's October 2025 research, The Widening AI Value Gap, found that 60% of companies are generating no material value from their AI investments – and that the gap between leaders and laggards is widening year on year, not closing.So, the question facing an Indian enterprise boardroom in 2026 is not whether to spend more on AI. It’s why so much already spent has given so little. The answer doesn’t lie in the model. It lies in the infrastructure beneath it. Most enterprises do not have an AI problem. They have an infrastructure problem that AI has exposed. Data sits scattered across systems that were never designed to talk to each other. Workflows depend on manual hand offers nobody has bothered to map. When you layer an AI agent on top of that environment, you have not accelerated the business. You have automated the chaos.The pattern separating the enterprises pulling ahead from the ones generating impressive demos is consistent: Fix the foundation, then add the intelligence. The leaders are not deploying smarter models. They are doing the unglamorous work first. They’re cleaning the data, consolidating channels, defining workflows, and deciding which tasks AI can take and which still need human intervention. By the time the model arrives, the inputs are clean and the success metric is already defined. Secondly, they apply AI where it accelerates execution, not where it adds complexity. AI works best when the inputs are structured and the output is measurable—a resolution time, a conversion rate, an order cost. As KPMG's Global AI Pulse for Q1 2026 puts it across in its India cut: "Enterprises are investing with confidence, but execution is now the real differentiator." That is the right framing. Execution, not experimentation.Third, they treat AI adoption as a business discipline, not a technology project. The work is embedded in frontline operations, owned by the business line, and measured against a number the CFO already cares about. This matters in any market, but it matters acutely in India, where enterprises cite regulatory, compliance complexity, and organisational resistance as their biggest AI challenges. Those challenges aren’t resolved by a procurement decision. They get solved with organisational discipline.This is the Indian agile enterprise moment. Companies with between 500 and 20,000 employees, large enough to feel the complexity tax, small enough to fix it, are where AI's productivity promise will be proven or broken. Recent mid-market research on the cost of complexity puts a number on the stakes: Indian firms lose roughly 27% of their AI budgets to complexity overhead – fragmented systems, overlapping tools, and integration drag – even as IT leaders report rising workloads from managing AI across that sprawl. India leads the global mid‑market on AI integration. It’s also paying the highest complexity tax to do it. These companies cannot afford to let a quarter of their software spend leak into integration debt. They do not need to. The leaders in this cohort will not be the ones who buy the most models. They will be the ones who rebuild their operations first and then let AI amplify the business they already run.India has the adoption lead. The question is who will turn it into the execution lead. That race is wide open, and it will not be won in the demo. It will be won inside the operation where mid-market IT leaders cut the cost of complexity, simplify the stack, and let AI do what it’s actually good at: Making a business that works, work better. (The views expressed are personal)This article is authored by Dennis Woodside, CEO, Freshworks.