Run a 15-Minute Weekly AI Briefing for Your Team
In most teams, AI knowledge is wildly uneven: one enthusiast who reads everything, a majority who catch fragments, and a few who've quietly given up on keeping track. The enthusiast's knowledge helps nobody, because it lives in one head.
The cheapest fix in organisational learning: a 15-minute weekly AI briefing. One short slot, one simple agenda, and the whole team's baseline rises together. Here's the format that works, honed from the training sessions we run with companies across the region.
The Agenda (15 Minutes, Strictly)
Minutes 1–5: Three things that happened
The briefer shares the week's three most relevant AI developments — relevant to your company, not to Twitter. One sentence of what happened, one of why it matters here. No slides. Slides turn a briefing into a meeting, and meetings die.
Minutes 5–12: One thing someone tried
The heart of the ritual. One team member shows something they actually did with AI that week — a prompt that worked, an automation that saved an hour, a failure worth avoiding. Two rules: it must be real (shown, not described), and failures count double. "I tried having it draft client emails and here's where it embarrassed me" teaches more than any success story.
Minutes 12–15: One question, one decision
Surface one open question ("can we put client data in this tool?") and make one micro-decision ("Priya investigates, answers next week"). Questions with owners get answered; questions without owners get repeated for a year.
Who Runs It
Rotate it. The permanent-expert model recreates the original problem — one informed person, everyone else spectating. Rotation means everyone must, occasionally, spend ten minutes sourcing three items — which quietly trains the whole team in the filtering skill itself. The rotation schedule is the curriculum.
Sourcing takes five minutes with a curated page: the briefer scans Cocoon's daily AI news (or their own equivalent), applies the five signal filters, and picks three. No research project required.
Why This Works When Training Days Don't
- Frequency beats depth for a moving target. AI shifts weekly; annual training is archaeology by month three. Fifteen minutes weekly compounds into the most current team in your industry.
- It normalises being mid-learning. When the senior manager shows their failed prompt, the permission structure of the whole team changes. That psychological shift outvalues any content.
- It creates a decision log. Fifty-two micro-decisions a year — tools evaluated, policies clarified, experiments assigned — is more AI governance than most companies manage with committees.
Starting Next Week
Don't ask permission for a new meeting; append it to one that exists. "Last 15 minutes of Monday's stand-up, we're trying an AI briefing for a month." Send the agenda above, volunteer to brief first, put the news page and this post in the invite.
By week four, expect the tell-tale sign it's working: items arriving mid-week — "saw this, saving it for the briefing". At that point the team has a shared radar, and you built it for fifteen minutes a week. For the personal version of the habit, start with How to Stay on Top of AI News in 10 Minutes a Day — and for what to do when the briefing surfaces a skills gap, our 2026 skills guide maps the priorities.
Cocoon builds free AI tools and runs practical AI training for professionals and teams across Sri Lanka and Southeast Asia. Try the free tool from this article or talk to us about training.