The year 2025 represented a pivotal moment for enterprise AI, as innovative tools garnered global interest. Business leaders across cities, including Mumbai, were making substantial investments in artificial intelligence to enhance productivity and optimize operations across various sectors. In India, the enterprise AI landscape buzzed with initiatives like hackathons, proofs-of-concept, and ambitious AI-first strategies.
However, amid this enthusiasm, a more profound crisis was surfacing. Despite significant investments, most AI initiatives failed to transition from the lab to practical application. A decade later, with AI-native workplaces now commonplace, it’s pertinent to examine what went wrong and what successful organizations did differently.
Jagdish Ramaswamy, Digital Transformation Advisor and former president and Chief Digital and Information Officer at Hindalco Industries Limited, notes, “I see very few enterprises that have successfully operationalized AI. It is considered successful when businesses begin to see tangible benefits. Three key factors differentiate those who succeed from those who do not: identifying the right use case for AI, developing a robust data pipeline, and having AI-skilled personnel ready to collaborate.”
Research highlights the scale of the challenges organizations faced. A 2025 Forrester report indicated that only 10–15% of AI pilots evolve into sustained production use, with more than 60% failing to scale beyond controlled environments. Additionally, an IDC report revealed that only 4 out of 33 AI prototypes reach production, reflecting an 88% failure rate. Many enterprises launch multiple AI experiments, yet nearly half are abandoned before production due to high costs, unclear ROI, and a lack of operational capability.
Rajendra Deshpande, technology consultant and former CIO of Intelenet Global Services, observes, “When examining laggards, it’s clear they lack a coherent business strategy. They often conduct isolated proof-of-concepts without an overarching enterprise roadmap. Their data management practices are typically siloed, unclean, and poorly governed. Furthermore, the structure of the AI team tends to be fragmented, and their infrastructure remains more suited for research rather than production. Governance mechanisms are either lacking or reactive, and leadership continues to view AI merely as an IT project.”
Successful organizations, however, prioritize clarity on the problem at hand rather than focusing solely on technological capability. AI initiatives must be driven from the top as structured, disciplined plans designed to foster collaboration across different teams. Treating data as a vital asset and employing various techniques to leverage it effectively is crucial.
Deshpande adds, “A culture of change must permeate across functions, and it is essential for organizations to establish feedback loops to learn from failures.”
Moreover, a growing disconnect exists between AI adoption and quantifiable business impact. While numerous organizations experiment with AI, only a small fraction manage to convert these initiatives into measurable outcomes. The primary challenge for industries is not access to AI technologies, but rather the ability to operationalize them. Transitioning from pilots to full-scale production necessitates a shift from experimentation to execution, grounded in data maturity, alignment with business objectives, and scalable infrastructures.
For further information about ET Workplace 2035, visit https://workplace.economictimes.indiatimes.com/workplace2035.
Published on May 12, 2026, at 08:00 AM IST.
