Fewer than one in five enterprises extract transformative value from AI deployments, MIT Technology Review Insights poll finds
A custom-content report by MIT Technology Review Insights, produced in association with Atlan, finds that 16% of organizations report transformative returns from AI, while 23% see minimal or no returns.
1 source · cross-referenced
- A custom-content survey by MIT Technology Review Insights, in association with Atlan, finds that 16% of organizations report transformative returns from AI deployments.
- 40% of respondents report moderate gains in specific tasks or teams, while 21% report meaningful gains.
- 38% of enterprises are still piloting AI use cases and 35% use AI to assist employees via tools like co-pilots or chat assistants.
- Only 12% have AI agents in production and 3% say AI does most of the work in key processes.
- Experts cite a lack of enterprise context as a key reason organizations struggle to realize AI’s potential.
A custom-content survey by MIT Technology Review Insights, produced in association with Atlan, finds that fewer than one in five organizations (16%) are extracting transformative value from their AI deployments. The largest share of respondents (40%) report moderate gains limited to specific tasks or teams, while 21% report meaningful gains. Another 23% say they are seeing minimal returns or none at all.
According to the poll, 38% of enterprises are still piloting AI use cases and 35% use AI to assist employees through tools such as co-pilots or chat assistants. Only 12% report having AI agents in production, and 3% say AI is doing most of the work in key processes.
Prukalpa Sankar, co-founder of Atlan, argues that the missing layer is context: a context layer that sits between an organization’s business systems and its AI platforms can translate enterprise intelligence into governed, machine-usable context for AI agents or models. Without such a layer, Sankar says, AI systems risk operating with limited understanding of the business, leading to misaligned or ineffective outcomes.
The report frames the challenge as a gap between what models are technically capable of and how useful they are inside organizations today. It suggests that establishing a company-wide "context layer" could improve accuracy, reduce costs, and enable more autonomous agents by grounding AI in enterprise knowledge.
- Nov 10, 2026 · Wired
Kora, a Linux-based desktop OS from Cyanogen cofounder, targets AI power users with local model integration
Trust76 - Nov 10, 2026 · One Useful Thing — Ethan Mollick
Mollick argues new AI models approach "jagged AGI" across coding, business, and humanities
Trust78 - Nov 9, 2026 · The Verge — AI
Former DOJ antitrust chief launches AI legal-research startup to challenge monopolies
Trust72