THE BUSINESS CHIEF INTERVIEW

THAT IS THE MEASURABLE RETURN on 95 % of corporate generative AI pilots, according to MIT’ s NANDA initiative, and it is a strange number to set beside the other one: corporations have spent between US $ 30bn and US $ 40bn arriving at it. In most large companies, there is now a piece of software nobody has opened in a year and nobody has cancelled either. Gartner expects more than 40 % of agentic AI projects to be cancelled by the end of 2027 – which is to say the cancelling has barely started. In an interview with Business Chief, Praveen Prabhakaran, Chief Operating Officer of UST, argues that the issue is often not the underlying technology but the way companies define and prepare pilots. After 24 years in technology and financial services, including India’ s first repo and forex platform, he says many organisations are evaluating AI outputs without establishing the business foundations needed to produce measurable outcomes. Today, he runs the firm’ s strategic verticals across life sciences, healthcare and financial services, with the UK and Europe under his wing as well.“ The word pilot has been misconstrued in the AI world,” he tells Business Chief.“ In the AI world, people are just playing on an output instead of an outcome. The base definition itself is wrong.

People tend to see that the pilots don’ t work. In the first place, they were never doing a pilot.” A proper pilot, in the digital era, was a whole product exercise. Feature engineering, market feedback, data pipelines, all of it.

Four building blocks nobody should outsource It’ s worth pausing on the phrase:‘ AI agent.’ Where did it come from? Who invented it? For all the mystique, an AI agent is a chatbot with a job and the run of your systems, your calendar, your codebase, a browser free to take a series of steps on its own and often disinclined to ask first. Nearly every deployment that works today is writing and reviewing code, fielding support tickets and retrieving and reconciling documents. Writing about automation in Science in 1960, Norbert Wiener reached for the Sorcerer’ s Apprentice and the Monkey’ s Paw, stories where the machine does exactly what it is told and ruins the one who asked. What Norbert framed as a warning is now the problem companies like UST are hired to solve. Praveen’ s own reason for being at UST is unusually unguarded for a COO. What drew him, he says, was the founders’ three-line philosophy, relevance for customers, significance for colleagues,

businesschief. com 59