One customer question, one shared learning record
Start with one active customer question. Pull the related Jira or GitHub work into the repo, capture what the team learns, and sync the useful context back before the decision drifts from the evidence.
Bring the evidence local
The team sees the issue, customer context, comments, attachments, and current state in one place instead of rebuilding the problem from browser tabs and memory.
Capture what changed
Record the hypothesis, customer signal, tradeoff, evidence, and next step so product, engineering, and AI sessions work from the same current understanding.
Push the learning back
Jira or GitHub stays visible to stakeholders, but the provider gets updated with the learning that should shape the next product decision.
The product problem is learning too late
Customer context gets separated from the work
The signal that changed the team's mind lands in a call, a chat thread, a document, or an AI transcript, while the work keeps moving somewhere else.
Feedback loops get too long
Teams learn what the customer really needed after the plan is already expensive to change.
Status hides the learning gap
Leaders can see progress, but not whether the team has enough evidence to trust the next product decision.
Start with one customer-learning loop
Run this as a small operating experiment, not a rollout. Pick one active customer question, let Jira or GitHub keep the audit trail the business expects, and use imdone to keep the question, evidence, decision, and next step close to the work.
Hypothesis
If one team keeps customer signals and build evidence in the same working context, they will make a better product decision before implementation momentum hides the uncertainty.
Setup
Pull the related Jira or GitHub work local, capture the customer question, evidence, decision, and next step, then push the useful learning back to the provider the team already trusts.
Signals to watch
Track how quickly customer feedback changes the plan, what evidence supports the decision, and whether the team avoids building from an assumption nobody has checked.
Plan the first customer-learning loop before the next assumption hardens
Start with one real workflow example, then decide where the team should keep customer evidence with the work.