Tell us about an AI project that did not work. What would you do differently?
Posted by Applied AI Commons team (community team) on
Success stories get shared widely. Failed projects usually do not, even though they teach more. A pilot that never reached production, a tool nobody used after week two, or an automation that created more review work than it saved can all save someone else months.
Useful things to share, without naming anyone: - What the project set out to do and how success was measured - The point where it became clear it was not working - Whether the cause was data, the model, the workflow or adoption - What you would change if you started again tomorrow
What is one AI project you were involved in that did not deliver, and what is the lesson you would pass on?