How It Works
How AI workflows can learn from experience.
Follow a structured approach to carrying relevant task context and decision rationale into related AI-assisted work.
How it works
From task experience to a reusable method.
Capture what mattered, structure it for reuse, then review and improve it as work continues.
- 01
Capture
Identify useful context, instructions, methods, and decisions from a task.
Why it matters: The details surrounding the work explain what made the approach relevant.
- 02
Structure
Organize objectives, assumptions, constraints, evidence, and rationale into a clear experience.
Why it matters: A consistent structure makes an experience easier to review and apply appropriately.
- 03
Apply
Use the experience to guide a compatible AI workflow or a future task.
Why it matters: The experience informs work while people and systems make the current decision.
- 04
Evaluate
Review the actual outcome against the objective and the evidence available.
Why it matters: A result shows whether the approach was useful under those conditions.
- 05
Improve
Use reviewed evidence and lessons to revise the experience.
Why it matters: Changes can address gaps without losing the rationale behind earlier decisions.
- 06
Reuse
Make the revised experience available to subsequent compatible workflows.
Why it matters: A clear version and compatibility context help determine when it applies again.
Illustrative example
Refine an incident response with evidence.
A team records a diagnostic method together with the deployment conditions, logs, assumptions, and result. When a similar incident appears, an engineer checks whether the conditions match before applying the method.
Iteration 1
First review
A retry resolves one incident, but later review finds a configuration mismatch behind the failure.
Iteration 2
Evidence added
The team records the relevant deploy manifest, error traces, recovery checks, and the conditions that mattered.
Iteration 3
Method refined
The next version checks configuration before retrying and retains the earlier reasoning for comparison.
Continue exploring