Wipro CTO: Under 10% of Enterprise AI PoCs Scale - Why the ROI Gap Persists

Enterprises are failing to generate real value from artificial intelligence because the four conditions for technology adoption have not aligned, Wipro CTO Sandhya Arun said at the Nasscom ER&D Summit in Bengaluru on September 1.

Arun identified four levers enterprises need to adopt technology at scale: scale (solving real-world problems), advantage (a meaningful step change), trust (a technology users can confidently rely on), and adoption (working within real-world economics). She said AI is not yet meeting all four.

Key takeaways:

• Fewer than 10% of AI proof of concepts (PoCs) are able to scale from lab to production
• Enterprise AI adoption has lagged due to change management friction, difficulty orchestrating workflows, high upfront investment, unclear ROI, and rising token costs
• Cognizant CEO Ravi Kumar identified the adoption gap as high token consumption without linkage to return on investment
• Large language model providers are responding by embedding AI directly into enterprise workflows to accelerate revenue realization and justify R&D spending
• Trust remains a barrier: employees fear AI will replace them rather than augment their work
• Arun raised ethical concerns about AI training use cases that record workers' physical movements without clear consent frameworks

The comments came as enterprise IT spending on AI continues to rise sharply, with organizations investing heavily in AI infrastructure while struggling to move from demonstration to deployment at scale.

Read the full article on Business Standard