Review an analytics project with AI
Compare a completed analysis with the original plan, identify repeatable lessons and assign concrete improvements for the next project.
Review an analytics project with AI after delivery, when the final chart no longer shows the three rounds of rework behind it. A change in the definition of an active customer may have caused the biggest delay. The analysis-retrospective skill helps a team reconstruct what happened and decide which change would help the next analyst.
Compare the original brief with actual timings, decisions and outputs. Check whether an apparent cause was simply the last visible symptom. Some lessons belong in a reusable template, while others apply only to a particular data source or team.
What it does
- Identify differences between the plan and actual delivery.
- Record what worked well enough to repeat.
- Trace a recurring problem to an evidenced cause.
- Turn a lesson into a specific checklist or kickoff change.
- Assign an owner and a follow-up date to each agreed action.
Install and use
- Give the agent an authorised completed project, original brief and relevant participant accounts. The retrospective can be drafted without an external analytics account.
- Ask, “Why did we redefine the customer metric during final review? Show the relevant decisions and suggest one change for the next kickoff.”
- Review the findings with the people involved and decide how the team will check whether the change helped.
FAQ
Source
Made by Nimrod Fisher. Original skill: analysis-retrospective, published on GitHub under the MIT licence.
View on GitHub