A team decision log for AI coding sessions
A team decision log captures the decisions your team makes during AI coding sessions, along with the reasoning behind them, automatically. Instead of decisions disappearing when a session ends, they become searchable team memory that future sessions recall, so the team never re-litigates a settled choice.
Why decisions get lost
Decisions get lost because they are made inside AI sessions that vanish when they end. The choice to use one library over another, the reasoning, the trade-offs considered, all of it lives in a session transcript no one will ever reopen.
Months later, someone asks why the team did it this way, and the answer is gone. The decision gets re-opened and re-argued from scratch, often landing somewhere different and inconsistent.
How a decision log builds itself
A decision log builds itself when a memory layer captures decisions from each session as they happen. It records what was decided and why, then stores it as searchable team memory, with no one stopping to write a record.
Because it is captured automatically, the log is complete rather than the handful of decisions someone remembered to document. And because it is searchable, the answer to why did we do it this way is one query away.
What gets recorded
What gets recorded is the decision, the reasoning, and the alternatives that were ruled out. The next time the topic comes up, the team's AI surfaces the prior decision instead of starting a fresh debate, so settled choices stay settled.
Decisions made out loud in meetings and written in documents flow into the same log, so the record is the whole team's reasoning, not just what happened in the editor.
When decisions change: reconciliation and Dreaming
Decisions are not static, and the log should not pretend they are. Teams revise choices as they learn: a library gets swapped, an approach gets reversed, a convention gets tightened. When two sessions record different decisions on the same topic, the memory has to know which one holds now, not just keep both side by side.
Wemory reconciles this with a daily curation pass called Dreaming. It reviews the memory, groups the decisions that touch the same topic, promotes the ones that have settled into verified team knowledge, and detects genuine conflicts between them. Most of the time it resolves the picture on its own, so the decision your AI recalls is the one that holds now, and the memory evolves as the team does.
When Dreaming cannot tell which decision should win, it does not guess. It surfaces the conflict as a question for a human on the team to settle, and that call becomes the record. So the log keeps up with how the team's thinking changes, with a person in the loop exactly when there is real doubt.
Frequently asked questions
What happens when two decisions conflict?+
Wemory's daily curation pass, Dreaming, reconciles decisions on the same topic: it promotes settled ones into verified knowledge and detects genuine conflicts. When it cannot resolve a conflict on its own, it asks a human on the team to make the call, and that decision becomes the record your AI recalls. The log evolves as your decisions do.
Is the decision log automatic or do we maintain it?+
It is automatic. Decisions and their reasoning are captured from your AI sessions, meetings, and documents as work happens, so the log stays complete and current without anyone maintaining it by hand. Reconciling changed or conflicting decisions is handled by the daily Dreaming pass, with a human stepping in only for genuine ambiguity.
Can we search past decisions?+
Yes. The decision log is searchable team memory, so you can ask why a choice was made and get the decision, its reasoning, and the alternatives considered, surfaced from whichever session or meeting it came from.
How is this different from commit messages or a wiki?+
Commit messages and wikis only hold what people stop to write, which misses most of the reasoning. A decision log captures decisions automatically from the sessions and meetings where they actually happen, including the why and the rejected alternatives.
What does it cost?+
Wemory is 20 USD per seat per month, first month free, no card required. Seats are prorated with no commitment, so the cost tracks your team size.