























Manish Prasad, President & MD, SAP Indian Subcontinent.
Boardroom conversations for many enterprises has moved from AI-readiness to the cost and business value of the technology, said Manish Prasad, President & MD, SAP Indian Subcontinent.
During SAP’s flagship SAP NOW AI Tour India, officials noted that nearly 67 per cent of Indian businesses are moving onto execution of AI projects rather than pilots in just 2 years. Yet, geopolitical volatility, rising compute costs, and cybersecurity pressures are making technology more expensive. In such circumstances, approximately 90 per cent accuracy of consumer grade AI (let alone AI for enterprises) is insufficient for finance or supply chain decisions.
“Two years back, Boards and CEOs were waiting for use cases for AI. Now, they are questioning the economics of technology and the relevance of that, from a use case standpoint in the context of their business,” said Prasad.
However, rather than only focusing on the cost impact, he pushed for companies to think about the real agentic value for enterprises from agents collaborating with each other and other human employees to improve the KPI of a business.
“The true agentic piece starts coming in when agents start to collaborate with each other to further improvise the KPI of a business or with respect to the governance. That’s the state in which we are getting into from an agentic layer standpoint. We probably still don’t know the end state,” said Prasad.
Despite this need for AI collaboration and thus a need for AI-ready skills, almost 8 in 10 organisations are not convinced their upskilling can keep up with AI advancements, as per SAP’s Value of AI Report 2026. Even in terms of AI governance, businesses say they’re either not fully ready when it comes to skills (11 per cent) or processes (14 per cent), despite AI governance being a key focus.
SAP found that 78 per cent believe data availability and quality are critical to AI-driven decision-making but 67 per cent enterprises struggle with low-quality data. Considering this, Prasad argued that sovereignty may not yet be required for every organisation in every industry,
“There’s enough level of security and reliability with the data residency happening in the country as such. And that’s true with every technology company. As you move up and look at end to end sovereignty, it is customer specific. So if you look at defense, if you look at national critical infrastructures, railways, each one of them have different levels of sovereignty expectation from a technology standpoint. It gets developed for each one of them,” he said.
Published on June 14, 2026
此内容由惯性聚合(RSS阅读器)自动聚合整理,仅供阅读参考。 原文来自 — 版权归原作者所有。