Trust will define AI’s next phase, says Vishal Sikka, flags LLM-enterprise gap

Vishal Sikka said the biggest opportunity in artificial intelligence lies not in building ever-larger language models but in bridging the gap between LLMs and real-world enterprise users.Speaking at the India AI Impact Summit 2026 on February 19, the founder and chief executive of Vianai Systems said AI can deliver dramatic productivity gains, but its impact remains uneven within organisations.“People who understand how to use AI are astonishingly effective with it,” Sikka said, citing cases where decisions that earlier took months or even a year were completed in days using AI-driven simulations. However, he described the current state of AI as a “jagged frontier”, where progress is powerful but inconsistent.Sikka said enterprises face a wide gap between what large language models can technically do and how value is delivered on the ground. “Overcoming that gap is where a lot of value-creating opportunities lie,” he said, adding that without correctness and trust, AI adoption will remain constrained.He also flagged persistent risks, including hallucinations, safety concerns and high energy consumption, arguing that India should focus not only on adopting current AI systems but on building safer and more efficient ones.Also Read: OpenAI deepens India push with Mumbai, Bengaluru offices and campus AI driveHis remarks echo comments from Nandan Nilekani, who recently said enterprise execution, rather than opportunity, is the main risk in AI adoption, as well as Ravi Kumar S, who noted that AI orchestration within companies will not happen “magically”.Together, the comments reflect a growing consensus among industry leaders that while LLMs have unlocked unprecedented capabilities, the harder challenge now is making AI work reliably at scale inside enterprises shaped by legacy systems, regulation and real-world risk.Also Read: How serious is AI risk to Indian IT services companies? Vishal Sikka explains

Source link