95% organisations get zero return from using AI tools, MIT study reveals
Cambridge: Despite global companies rushing to integrate artificial intelligence (AI) systems to cut costs and drive profits, a recent study from the Massachusetts Institute of Technology (MIT) has revealed that most of these initiatives have failed to deliver measurable results.
The report, titled The GenAI Divide: State of AI in Business 2025 , cited by NDTV on Monday, found that 95% of organisations that implemented AI systems have seen no return on their investments. Despite an estimated $30-40 billion in enterprise investment into generative AI (GenAI), the study uncovered that the vast majority of AI projects have had little to no impact on profit and loss (P&L) performance.
The research surveyed 300 AI deployments and gathered insights from approximately 350 employees. Among the most commonly adopted AI tools were ChatGPT and Copilot, but only 5% of AI pilots have managed to extract substantial value, while most others have stagnated without any measurable financial impact.
"Over 80 percent of organisations have explored or piloted them, and nearly 40 percent report deployment. But these tools primarily enhance individual productivity, not P&L performance. Meanwhile, enterprise-grade systems, custom or vendor-sold, are being quietly rejected," the report stated.
The study also pointed out that the failures were not necessarily due to AI models not functioning efficiently. Rather, the challenges arose from difficulties in integrating these systems with existing company workflows. Furthermore, many companies face a "learning gap" within their workforce, complicating the adoption of new AI tools. Company executives, however, have reportedly blamed AI models' performance as the root cause of the lack of returns.
In a related example, Dane Mathews, Chief Digital and Technology Officer at Taco Bell, explained that the fast-food chain has slowed down its AI rollout at drive-through restaurants. Despite expectations, AI technology proved counterintuitive, with Mathews acknowledging that human workers were often better suited to take orders, particularly during busy periods.
This echoes findings from a June study by Apple, which cast doubt on the capabilities of current AI models. In its report titled The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity , Apple argued that AI systems like Claude, DeepSeek-R1, and o3-mini do not actually reason as humans do. According to Apple, these models excel at pattern recognition but struggle when confronted with complex or altered questions, often falling apart when patterns become too complicated.
Let the Truth be known. If you read VB and like VB, please be a VB Supporter and Help us deliver the Truth to one and all.
KKR defeat LSG via Super Over
Lucknow (PTI): Kolkata Knight Riders edged Lucknow Super Giants via Super Overs in a battle between two bottom-placed teams in the IPL, here on Sunday.
Chasing a modest 156, LSG suffered a batting collapse and managed to tie the contest and take it to Super Over with No. 9 Mohammed Shami striking a last-ball six against a wayward Kartik Tyagi, who leaked 16 runs in the final over.
But Sunil Narine bowled a stunning Super Over conceding just one run and taking two wickets to give KKR an easy target.
Rinku Singh then finished the chase with a boundary off first ball from Prince Yadav.
Earlier the KKR were in deep trouble with 93/7 in 15 overs but Rinku smashed a sensational 83 not out from 51 balls including four sixes in a row in the final over to lift them to 155/7.
Cameron Green (34) was the only other batter to reach double-digit scores as KKR suffered a familiar batting meltdown with Mohsin taking his maiden fifer.
In reply, LSG continued their dismal show with the bat to succumb to their fifth loss in a row.
Brief Scores:
Kolkata Knight Riders 155/7; 20 overs (Rinku Singh 83 not out, Cameron Green 34; Mohsin Khan 5/23). Lucknow Super Giants 155/8; 20 overs (Rishabh Pant 42). KKR won via Super Over.
