TL;DR
AI shortcuts productivity early, but it also hides fragile practice. Most teams hit a competence cliff 6–12 weeks in: outputs multiply, standards diverge, and the organization’s implicit skills fail under real work. The fix isn’t a bigger model or more slides — it’s a deliberate program of contextual practice, rapid calibration, and bounded autonomy. This issue gives a crisp mental model, the five failure modes that create the cliff, and a practical 30/60/90 plan leaders can run this week to keep progress compounding.
A quick note
Amazon just announced 30,000 layoffs. UPS cut 48,000. Both cited AI as a key driver, and it’s not stopping there.
If you were recently affected and want to pivot into an AI-related role, I want to help. I’m putting together a small effort to connect people with resources, mentors, and companies hiring for AI-skilled roles.
I’ve been in the AI space for 8+ years, worked with ML systems at Meta, founded an AI education non-profit that reached 70,000 people, and now run an AI testing platform where I see firsthand how companies are implementing AI and reshaping their approach to business.
If that sounds useful, you can fill out the form below. I’ll share what I learn as I help people navigate this shift.
Why teams forget (short)
People are not the problem; systems are.
Knowledge has a half-life. Without reinforcement, most insights evaporate.
Docs are passive; people need prompts and stories.
Learning divorced from daily work doesn’t stick — it’s information, not capability.
Fix those three, and you stop relitigating the same decisions three times.
The memory-first stack (high level)
Treat organizational memory like a product. Build these layers in order:
Capture (automatic + manual): short decision notes, incident summaries, outcome snippets. Capture when memory is fresh.
Store (stories, not manuals): 200–400 word narratives that explain what happened, why, and what changed. Tag them. Link artifacts.
Attach (contextual links): surface those stories inside tickets, meeting agendas, onboarding checklists. Make them discoverable where decisions are made.
Surface (recall & nudges): daily/weekly digests, PR reminders, “related story” suggestions.
Practice (apply): micro-tasks that force immediate repetition — shadowing, small experiments, reproduction tasks.
Measure (metrics): recall rate, recurrence rate, playbook adoption, time-to-resolution.
Start with capture + store + attach. The rest scales from there.
Tactical playbook — what to do in the next 30/60/90 days
0–30 days — stop the leak
Mandate: every meaningful change (product, policy, infra, process) gets a one-paragraph Decision Note linked to its ticket or PR. No exceptions.
Meeting change: add a 10–15 minute “one story” slot to weekly team meetings — one thing learned, one action to test.
Template (copy-paste):
DECISION NOTE
Title: <short>
Context: <one sentence>
Decision: <what we did>
Expected outcome: <metric>
Actual outcome: <filled later>
Owner / Date
30–60 days — build the storybank
Convert 8–12 recent Decision Notes into story entries. Keep them short; tag by theme and outcome.
Assign an owner for the storybank (rotating editor is fine). Enforce linking: every story must link to at least one ticket, PR, or dashboard.
Run one “story conversion” hour with each team: transform postmortems into stories.
60–90 days — surface & practice
Add one recall hook: a) a “related stories” section in the PR template, or b) a weekly digest that maps stories to current initiatives.
Create 3 micro-practice tasks derived from top stories for onboarding and new projects.
Run a cross-team “story hour” — each team shares one win and one failure and the playbook it produced.
Turn postmortems into playbooks — exact steps
Postmortems end with “lessons learned.” Make them actionable playbooks:
Extract 3 reproducible steps that someone can follow next time.
Define a trigger (when to run this playbook).
Assign an owner and a test case that proves the playbook works (replay a past incident in a sandbox).
Publish the playbook, link it to the story, and add it to onboarding tasks.
Playbook mini-template
PLAYBOOK: <name>
TRIGGER: <condition>
STEP 1: <action>
STEP 2: <action>
STEP 3: <action>
OWNER: <name>
TEST: <how we’ll validate this works>
Low-friction recall experiments (non-engineering)
You don’t need heavy tooling to start surfacing memory.
Meeting nudges: create a 2-line “Did we do this before?” reminder in agendas that links to one story.
Email digest: weekly 3-line roundup — 1 new story, 1 playbook, 1 micro-action.
Onboarding task: every new hire completes two story-based micro-tasks in week 1.
Office hours: a 30-minute “story clinic” where people bring a decision and someone shows a relevant past story.
These make memory visible without new systems.
Metrics that matter (pick 3)
Measure impact, not activity.
Recall rate: % of new tickets/PRs referencing a relevant story (goal: 50–60% in 90 days).
Playbook adoption: % of incidents where an existing playbook was used.
Reoccurrence rate: rate of repeated failures in the same class (should trend down).
Time-to-resolution (recurring issues): median time to fix problems that have happened before.
Pick three and report them weekly — ownership creates urgency.
One concrete example (short)
Problem: Support keeps triaging the same billing edge-case.
System fix:
Capture the last incident as a 200-word story linked to the ticket and PR.
Extract a 3-step billing playbook with a clear trigger.
Add the playbook task to new support hire onboarding and the PR template for billing changes.
Result: repeated tickets drop, onboarding time for billing competence drops, support morale improves.
Common traps (and how to avoid them)
Trap: “We documented it, so we’re done.” → Reality: documentation without recall is shelfware. Pair capture with surfacing.
Trap: “We’ll automate later.” → Reality: capture now while memory is fresh; automation can index later.
Trap: “Tools will fix this.” → Reality: tools without rules = clutter; rules without tools = friction. You need both. Start with rules.
Quick playbook for founders & leaders (1 page)
This week: add Decision Note to PRs/incidents. Run a 15-minute learning slot in next team sync.
This month: convert 8 recent decisions into stories and tag them. Assign a storybank editor.
In 60 days: require 1 related story per PR and add 3 story-based onboarding tasks.
Quarterly: measure recall rate and reoccurrence rate; iterate.
Two copy-paste prompts for your teams
For PMs / Team leads (before a kickoff)
Before kickoff: write a 2-line “what could go wrong” premortem and link to any relevant story. Add one measurable signal we’ll watch.
For anyone closing a ticket/PR
Before close: add a 1-paragraph Decision Note (Context / Decision / Expected outcome) and tag related stories.
Final note — culture × systems
Culture matters. But culture without systems is optimism. If you want your team to get better, design for memory: capture, attach, surface, practice, measure. Do that and learning stops being a nice-to-have — it becomes your compounding advantage.
👉 If you found this issue useful, share it with a teammate or founder navigating AI adoption.
And subscribe to AI Ready for weekly lessons on how leaders are making AI real at scale.
Until next time,
Haroon