What is a lessons learned log?

A lessons learned log is a list of what a team found out while doing the work, kept so it changes how the next piece of work is done. Each lesson answers three questions: what happened, what we learned, and what we will do differently. A project manager might collect them at a retrospective, a postmortem or the end of a project; a developer might write one down the moment a fix took a day longer than it should have.

Why most lessons are never applied

The usual lessons learned document is written once and read never. It sits in a folder nobody opens at the start of the next project. With AI assistants the problem is sharper: an assistant starts each session without the last one’s memory, so it repeats the mistake it made yesterday unless somebody tells it again.

What makes a lesson useful

• It is specific: "check the contract status as well as the order status" rather than "be more careful with data".
• It is linked to the work it came from, so anybody can read what actually happened.
• It says what to do differently, as a step someone can follow.
• It ends somewhere: applied as a standing instruction, or set aside with a reason.

How fenbs uses it

Every fenbs board has a Lessons learned page. A lesson gets its own number, LES-001, and links to the tasks it came from; when a task reaches Completed, the board asks "Anything we learned?". New lessons are part of the AI context every connected assistant reads before it starts, so they are applied without anybody remembering to mention them. A lesson worth keeping is applied: as a context note, or as a proposed rule that a person decides on the Decisions and rules page. Assistants can record lessons too, over MCP, signed with their name.