Summary
Retrieval and spaced follow-up ask learners to recall and reuse what they learned after time has passed, instead of only re-reading or re-watching materials.
For this project, this pattern helps keep AI workflows from becoming one-shot training events that never change durable behavior.
Evidence status
Assessment: evidence-backed
Primary support:
- Source - Improving Students' Learning With Effective Learning Techniques (Dunlosky et al., 2013)
- Source - Retrieval Practice Consistently Benefits Student Learning (Agarwal et al., 2021)
Why this pattern belongs here
- Passive review is a weak reinforcement strategy.
- AI enablement will fail to transfer if teams only attend a session and never recall or reuse the patterns later.
- Retrieval helps reveal what people actually retained, not just what felt familiar in the room.
What this pattern is trying to achieve
- improve retention
- strengthen later recall in real work
- surface weak learning before it becomes a delivery problem
- keep the training program lightweight but durable
When to use it
- after workshops
- after rollout demos
- after first workflow adoption
- during office hours or follow-up checkpoints
When not to misuse it
- do not reduce retrieval to trivia detached from the workflow
- do not overload people with long quizzes that feel academic rather than useful
- do not assume a single follow-up is enough
Patterns and practices
- short recall prompts a few days after training
- ask participants to name a workflow, guardrail, and verification step from memory
- use brief scenario questions instead of only factual recall
- revisit the same workflow later with a slightly different case
- ask teams to report where they applied the pattern in live work
Good forms for this project
- short Slack or Teams follow-up prompt
- office-hours checkpoint question
- lightweight scenario-based recall card
- manager or lead check-in using a small set of standard questions
Anti-patterns
- using retrieval only as compliance testing
- making follow-up so heavy that teams disengage
- relying on rereading slide decks as reinforcement
- separating recall from real application for too long
Example application in AI enablement
Three days after a workshop, ask participants to respond from memory to:
- when should this workflow stay in learning mode
- what would you verify before trusting the AI output
- where would you escalate to peer review
What should accompany this pattern
- a clear initial learning experience
- real-work application soon after
- a manager or peer support loop when possible