AI weekly routine for the founder — 90 minutes that compound

If you're a founder reading 5+ status updates a day and still feeling like AI is something the team is "trying" but you're not — the missing piece is a 90-minute Friday block on your own calendar. Not a strategy offsite. A weekly habit.

Why founders specifically need a weekly AI routine

The founder is the only person in a 30-500-person company who can see across functions weekly. AI doesn't change that — but a founder who treats AI as a once-a-quarter strategy item and not a weekly habit ends up with the BCG 2025 problem: deployed but not valuable. ~78% of orgs deploy AI; only ~25% see meaningful value.

The 90-minute routine is the smallest commitment that produces signal. Less and the routine evaporates; more and you crowd out the actual job.

Definition: Productivity dip — the period (typically day 8-15 of any new tool) when usage drops because the user is still slower than their pre-AI baseline. Past this dip, retention soars.

The 90-minute Friday routine

Block 1 (0-30 min): review what worked

Open your AI tool (sanctioned, logged, with org access). Review the 5 most-used prompts your team ran this week — most enterprise AI dashboards expose this. Ask three questions:

1. Which prompt produced the highest-leverage output (something a leader actually used)?
2. Which prompt got run 20+ times by different people — meaning it should become a shared template?
3. Which prompt produced something you wouldn't want a customer to see — and why did it ship?

Write a 5-line memo. That memo is the input to Block 3.

Block 2 (30-60 min): sharpen the prompt library

Open the shared prompt library (you have one — and if you don't, this is the week you make one). Pick the highest-leverage prompt from Block 1 and improve it. Add a section for "what to NOT include", add a worked example, and add a "what good looks like" snippet. Then pick one prompt that's been getting used wrong and either fix it or retire it.

Prompt library entry template:
- Purpose (one sentence)
- Inputs (what to paste)
- Output shape (JSON / markdown / table)
- What good looks like (1 example output)
- What to NOT include (privacy, scope creep)
- Owner (a real human's name)

Block 3 (60-90 min): plan next week

Decide three things: (a) which function gets the founder's AI focus next week (you can only push one at a time), (b) which prompt you'll personally test on Monday morning, (c) which leader gets the 5-line memo from Block 1 with one specific ask. Send the memo before you close the laptop. Don't save it for Monday.

Tool tip (Course for Business): In our 6-week program we run a weekly founder cadence alongside the team activation. The principle is Augment, don't replace — the founder doesn't write more memos, AI helps the founder produce one sharper memo per week. We've found that founders who hold the 90-minute Friday block for 4 consecutive weeks change company-level adoption visibly. See course.aiadvisoryboard.me/business.

What NOT to do in the 90 minutes

• Don't schedule meetings inside this block. The block dies the first time you let a customer call sneak in.
• Don't review every team's AI use. You can't. Pick one function per Friday on a rotation.
• Don't generate net-new strategy documents. This is review + sharpen + plan, not a writing session.
• Don't share the prompt library with your CEO peer group "just because". Internal first; external, never without thinking.

A 6-week founder rotation

Week 1: revenue (sales + marketing prompts). Week 2: ops (finance + people-ops prompts). Week 3: product (discovery + spec prompts). Week 4: engineering (Copilot + review prompts). Week 5: customer success (renewal + escalation prompts). Week 6: meta — review the 5 weeks, kill what didn't stick, double down on what did.

By week 6 the founder has touched every function once, has a sharpened library, and has a list of "what to amplify next quarter".

Team scan (what AI champions report after week 1)

• The founder ran the 90-minute block on Friday — even when the calendar tried to eat it.
• The shared prompt library has at least one new entry contributed by the founder personally.
• One leader received a 5-line memo with a specific ask that landed on Monday.
• 1-2 prompts that had been getting used wrong are corrected or retired.
• A "what to NOT include" line is added to the highest-leverage prompt.
• The team notices the founder is engaging with AI weekly — adoption nudges up without a memo.
• 1 prompt that nobody ran gets removed from the library; the library is shorter, not longer.
• The founder's personal Monday-morning AI test is on the calendar with a 15-minute block.
• A specific function is named for next week's focus before the founder closes the laptop.
• The 90 minutes finish on time; the founder doesn't blow into Saturday.

Tool tip (Course for Business): Founders we work with run with an AI Champions (1:15-20) ratio across the company — meaning a 60-person team has 3-4 champions, and the founder's Friday memo lands in the champions' Slack, not the all-hands. The 6-week program is structured around this cadence so the founder's habit and the team's habit reinforce each other from week 1. course.aiadvisoryboard.me/business.

Micro-case (what changes after 7-14 days)

A typical 90-person services-firm founder runs this routine for 4 consecutive Fridays. Week 1: discovers two leaders are using personal-account AI for client docs; brings them into the sanctioned tool. Week 2: kills three prompts nobody used; adds one "client-meeting prep" prompt that gets adopted in 5 days. Week 3: identifies a function (finance) that has zero AI adoption; pairs the CFO with a champion. Week 4: realizes the company prompt library is ~70% better than month start, with no extra headcount and no consultant. The founder closes Friday at 4:30pm with a clearer head than before AI, not despite it. By week 12, BCG-pattern productivity gains start showing in the leadership reports.

Note on this case: This example is illustrative — based on typical patterns we observe with companies of 30-500 employees, not a single named client. Specific numbers are rounded approximations of common ranges, not guarantees.

FAQ

Why Friday and not Monday?

Friday lets you review what just happened, while it's fresh. Monday's review is always stale. Plus a Friday plan-next-week doc lands in leaders' inboxes when they have time to think, not when they're already mid-sprint.

What if the company is too small for this — say, 30 people?

The routine still works; the rotation compresses. With 30 people you might do all 6 functions in 3 weeks instead of 6, since several functions overlap or share leaders.

What's the worst version of this routine?

Spending all 90 minutes generating new strategy memos with AI. That's not the routine. The routine is review + sharpen + plan, in that order, with a 5-line memo as the only output to others.

How does this differ from the COO routine?

The COO routine is daily and operational (Mon-Fri cadence). The founder routine is weekly and strategic. Different cadence, different purpose.

Does this overlap with your daily-management product?

Slightly. The daily-management OS surfaces what each team did, every day, at the company level. The 90-minute founder routine is what the founder does with that signal once a week. Complementary, not substitutes.

Conclusion

90 minutes a week. Reviewed, sharpened, planned. No new tools, no new headcount. The founders who hold this block for 6 consecutive weeks see compounding effects the rest of the company can feel. The ones who skip it stay in the BCG ~75% — deployed but not valuable.

If you want every employee in your company to ship their first AI automation in five days — book a 30-min call and we'll map your team's first week: course.aiadvisoryboard.me/business.