Companies question whether AI is delivering real business value

After three years of aggressive spending and experimentation, companies are now questioning whether artificial intelligence is delivering real business value. “The pixie dust phase is over. Everyone is now asking where the ROI is,” Didier Bonnet, professor of strategy and digital transformation at IMD Business School, told ET.

Speaking on the sidelines of CII Edge Summit,2025, Bonnet said many boards are struggling to make sense of AI investments, and CEOs are under pressure to show results. “Confused boards and overwhelmed CEOs” are already beginning to pull back as promised gains fail to materialise, he said. According to him, the problem is not the technology itself but how companies are using it. “Execution is the real issue, not AI.”

The problem, he explained, is that the majority of organisations have dozens of pilots without scaling them, spreading their efforts too thin. This approach, he said, follows the same pattern seen during earlier waves of digital transformation. Only a few sectors are doing AI well, with pharmaceuticals standing out. “They have clear, high-value use cases,” Bonnet said, pointing to areas such as drug discovery and clinical trials.


In many other industries, the gains from AI remain limited to individual productivity, such as faster writing, summaries or analysis. These improvements do not automatically translate into higher revenues or company-wide efficiency unless businesses redesign processes, workflows and roles, he noted.

He also questioned claims that AI is already leading to widespread job losses. Companies blaming AI for layoffs are often using it as a cover for weaker demand and fewer deals, he argued. “The real disruption is to the pyramid model in law, consulting and banking. If you cut entry-level talent, the whole operating model breaks.”

On global investment trends, he said the direction of AI will largely be shaped by the United States, where spending on computing power, data centres and research is concentrated among a small number of firms. This could either support long-term growth or turn into “one of the largest circular-investment bubbles ever”. A more likely outcome, he said, is a correction in valuations or a scaling back of announced investment plans that do not fully materialise.

For India, Bonnet said the biggest risk is remaining dependent on foreign technology. Without a shift from a services-led approach to building home-grown AI products, India could end up as “just a market” for US-developed tools. “If you don’t build your own capabilities, you don’t control your future,” he said. He also warned against fragmented investments across states, universities and data centres, especially as AI infrastructure is energy and water-intensive.

What gives India an edge, Bonnet said, is scale. With better coordination, the country can build AI systems suited to its languages and market needs. “India has the talent, the demand and the ability to do this differently,” he added. A clear national strategy and stronger alignment across government, industry and academia could turn AI into a long-term growth engine.