Why AI Transformations Fail

Most AI transformations stall for reasons that have little to do with the technology and everything to do with the people expected to make it work. 

In this Gamechanger Interview, Ajay Manissery Konchery, the CEO of Wisdom Tomorrow, recently appointed as the Co-Managing Director at Epic Group, a sustainable fashion and apparel manufacturing multinational company, joins Victor Filamor, Partner at Stanton Chase Greater China, to explain why so many AI investments fail to deliver a return. Drawing on Stanton Chase’s survey of 214 executives across 45 countries, he works through findings that should concern any board, returning again and again to why the problem is rarely the technology and almost always the people expected to make it work. 

Why Do Most AI Transformations Fail?

According to Ajay, the trouble starts with intent. Many companies cannot say precisely why they are deploying AI in the first place, whether the aim is a better customer experience, a more efficient supply chain, or some combination of the two. Without that clarity, the C-suite falls out of alignment; the technology gets launched first, and the strategy is retrofitted around a tool that is already running. Add weak data and inflated expectations of AI as a cure for long-standing problems, and the transformation stalls. 

 What Percentage of AI Projects Fail or Stall?

In the Stanton Chase survey of 214 executives, 77% had not moved AI past experimentation, so most organizations are still running pilots that have yet to prove their value. Ajay traces this to a lack of alignment before the work begins, since the CFO, the CIO, and the CEO often hold different expectations while the CHRO is rarely invited to weigh in. 

 Why Can’t Companies Measure the ROI of AI?

The survey found 65% of executives unable to show a measurable financial return on AI, and many said it was still too early to judge whether their spending had paid off. Ajay’s reading is that returns stay out of reach while AI sits in scattered pilots, since impact can only be measured once the technology moves into operations and the data underneath it is put in order. 

Who Should Own AI Transformation in a Company?

Ownership was scattered across the executives surveyed, spread between the CEO, the technology chiefs, and shared committees, with just 2% placing it with the CHRO. As one respondent noted, when responsibility is spread across a committee, no one is truly accountable. Ajay argues that because AI succeeds or fails on human factors, the people function belongs at the center of the decision rather than on its edge. 

How Can Leaders Build Workforce Trust in AI?

Trust sits at the center of Ajay’s argument. The reaction of employees, and of customers in consumer businesses, often decides whether a rollout works, yet leaders tend to discover this only once deployment is underway. Governance adds a second concern, since AI can carry bias and can lead a company astray without human judgment and supervision. There is also the tension between data privacy and transparency, and the risk of black boxes that no one in the C-suite can explain. Ajay’s counsel is to keep human oversight close, so that decision making does not run away from the people accountable for it. 

How Do You Make an AI Transformation Succeed?

Ajay’s answer starts with a clear reason for using AI, a pilot in one part of the business, and the data and people work done before scaling. To help leaders find where to begin, he built the AI Workforce Engine™ with Professor Joon Nak Choi, a tool that maps around 74 points where people and AI work together and produces a score that shows where the gaps sit and which areas to address first. The aim is a company where people amplify AI and AI amplifies people, rather than both dragging each other’s productivity down. 

Watch More

Read More

Watch the full interview here