MIT: 95% of enterprise GenAI pilots show little or no P&L impact
krishnan · x · 2026-10-22
- MIT NANDA's 2025 GenAI Divide report finds 95% of enterprise GenAI pilots show little or no P&L impact — integration, not the models, is the culprit; generic tools never learn the workflow and stall after the demo.
- Buying from specialized vendors or partnering succeeds 67% of the time; internal builds succeed about a third as often.
- Over half of GenAI budgets go to sales/marketing tools, yet the biggest returns show up in the back office: BPO, agency spend, ops.
- Based on 150 interviews, a 350-person survey, and 300 public deployments; directional. Suggested metric swap: cost per completed outcome, since seats and prompts never hit the P&L.
More from Companies & People
- Major US banks detect unauthorized third-party AI agents inside customer accounts, Axios reports — EthanJPerez · 2026-10-22
- Ben Bajarin: cybersecurity becomes the budget that lets enterprise AI scale — BenBajarin · 2026-10-22
- Tesla's Terafab aims to build one terawatt of silicon compute per year — yunta_tsai · 2026-10-22
- Anthropic accused of racing toward recursive self-improvement in public safety spat with Dario — wfithian · 2026-10-22
- Dev advocates vanish into frontier labs — community calls for them to build and share again — craigsdennis · 2026-10-22
- Quantinuum and Oracle open early access to Helios quantum computer via OCI Quantum AI Compute beta — AryHHAry · 2026-10-22