AI Use-Case Library Reveals 159 Real Deployments and Six Failures

{ "title": "AI Use-Case Library Launches With 159 Real-World Deployments, Six Failures, and a Warning for Budget Season", "body": "A free, searchable database of real-world AI deployments has landed, offering managers a chance to learn from 159 documented projects across 21 industries before they spend a dime. The AI Use-Case Library includes reported outcomes on 77 of those deployments and, crucially, six that were halted or reversed.\n\nThe library, published this week, requires no signup and lets users search by industry, function, vendor, or whether the project is still running. Every entry is linked to its source. The author, Alexis, writes that the six halted entries “might be the most useful entries in the file,” calling them “the cheapest lessons in the library.”\n\n## The Wins: Narrow Tasks, Measurable Results\n\nThe library documents successes that share a pattern: focused tasks with clear metrics. At OpenAI, 97.9% of employees now use its Codex agent, up from about 40% last August. The company’s Legal and Recruiting departments have adopted the tool. The UK’s National Health Service expanded its AI chest X-ray reading program with a £20M investment, helping over 4 million patients get faster lung-cancer diagnoses or all-clears. The system cuts analysis time on complex cases from eight days to four.\n\nIn food production, Wonder bought a bowl-making robot from Sweetgreen that produces 500 bowls per hour. A human line cook makes about 45 bowls per hour. Momenta’s urban Navigate-on-Autopilot software now runs in 680,000 to 900,000-plus production vehicles from Toyota, Mercedes-Benz, BYD, GM, and Audi. Pinterest’s AI-powered Performance+ campaigns drove the company’s first $1B quarter, accounting for about 30% of lower-funnel revenue. Advertisers who adopted Performance+ grew that spend at nearly twice the rate of those who didn’t.\n\nAt China Post’s Guangzhou hub, humanoid robots process up to 1,200 parcels per hour. The site moves 6.5 million pieces per day. “The wins are concrete and uneven,” the article notes. “Adoption is real where the task is narrow and the outcome is measurable.”\n\n## The Failures: When AI Backfires\n\nThe six halted deployments offer stark warnings. Ford rehired 350 quality inspectors after automated quality systems produced defects. Waymo paused robotaxis in four cities—Atlanta, San Antonio, Dallas, and Houston—after a robotaxi drove into an Atlanta flood and sat stuck for about an hour. The company also paused service in Dallas and Houston as a precaution for forecast weather. Meta dropped automated hate-speech moderation and switched to a Community Notes model. Within six months, abusive and racist posts targeting US legislators tripled.\n\nThe U.S. Department of Government Efficiency used ChatGPT to flag about $100M in National Endowment for the Humanities grants as “DEI-related.” The model labeled a Holocaust literature anthology that way. A 143-page court ruling found unconstitutional viewpoint discrimination in the AI-driven grant cuts. Wake County schools in North Carolina banned AI detectors after a student given a zero on a detector flag appealed and had the grade changed to 100 when a second teacher found no AI use. “The failures rhyme,” the article states, pointing to common patterns across Ford, Waymo, and Wake County.\n\n## The Audit Era: Budgets Under Scrutiny\n\nAmazon CTO Werner Vogels told Fortune that enterprises are shifting to cheaper open-source models because bills have become real. He cited Uber burning through its entire 2026 AI budget in four months, and another company running through half a billion dollars in a single month before capping employee usage. Bloomberg reported that OpenAI, Meta, and xAI are racing to undercut Anthropic on price as buyers scrutinize invoices. 404 Media documented Amazon, Adobe, Atlassian, and Citi throttling employee AI use; one firm’s monthly spend tripled past $15 million.\n\n“The AI market just moved from ‘can it do this’ to ‘what did it cost, and did it work,’” the article argues. “When budgets get audited, ‘we think it’ll work’ stops being enough.” The library is designed for that moment. “Precedent beats hype in a procurement conversation,” Alexis writes. The article advises users to filter by industry and function first, read halted entries before wins, and use linked sources in budget meetings.\n\n## What’s Next: Robots in Classrooms and Policy Debates\n\nThis fall, the Salamanca City Central School District in rural upstate New York, located on the Seneca Nation reservation, will deploy a Realbotix M-Series humanoid robot and an AI teacher’s assistant called Optio for about 500 high-school students. The deployment is part of the district’s own AI and robotics curriculum. The EdTech Innovation Hub called it “a landmark moment while noting it’s still a single-district test.”\n\nMeanwhile, the FTC opened public comment this week on a policy statement on AI accuracy. Knight Columbia published “AI as Social Technology,” arguing for treating AI as a social technology with civil-rights and regulatory implications. Anthropic committed $10M CAD to eight Canadian research institutions, including Mila, Vector, Amii, and the University of Toronto, providing Claude credits and startup access.\n\nThe article includes a poll for readers: “Twelve months from now, your organization’s AI spend is:” with options including “Bigger and clearly paying off,” “Bigger and still unclear if it pays off,” “Smaller capped by usage limits,” and “Flat mostly moved to open-source.” Last week’s poll of 334 voters asked about the AI trade 12 months from now. Results were evenly split: 25% said “Still booming and revenue catches up,” 24% said “A correction then a stronger market,” 25% said “A dot-com rerun,” and 25% said “Two markets: infrastructure cracks software holds.”\n\n## Related on Neura Market\n\n- AI Infrastructure & Hardware\n- Enterprise AI Adoption\n- AI Regulation & Policy" }