Before Implementing AI — See What Your Team Actually Does

When a founder of a 40-person manufacturing team told me they’d already bought AI licenses but saw no change in output, I realized they’d skipped the most important step: seeing what the team actually does.

How to see what your team actually does — before touching AI

Start with the founder’s decision: believe there’s value in seeing real work, not just hearing about it in meetings. Most founders assume they know how time is spent — until they look.

Step 1: Pick one team or function to observe for 5–7 days. Don’t announce it as an “AI project” — frame it as understanding current work to reduce overload.

Step 2: Have each person log their actual work in simple terms: what they did, how long it took, and whether it felt routine or complex. Use a shared doc or async update — no forms.

Step 3: At the end of each day, the founder reviews the logs and asks: Where did time go? What was repeated? What felt like firefighting? What was never on the plan?

This isn’t about surveillance. It’s about finding the 20% of routine work that eats 80% of time — the prime target for AI augmentation.

Tool tip (AiAdvisoryBoard.me):

Tool tip (AiAdvisoryBoard.me): The Plan → Fact → Gap lens turns vague complaints like “we’re always behind” into clear data: e.g., Plan was 2 hours/day on client reports, Fact was 4.5 hours due to manual data pulls from three systems. Gap: 2.5 hours of automatable routine. See how the 7-day diagnostic works.

What to look for in the Fact column

Not all work is equal. Focus on:

• Tasks done more than 3x/week by multiple people.
• Work that involves copying data between systems.
• Activities that feel “necessary but dull” — like formatting reports or chasing approvals.
• Anything that shows up in logs but never in meeting agendas.

These are your automation candidates — not the flashy, strategic projects that get discussed in leadership offsites.

Manager scan (2-minute digest example)

• Sales lead: Planned 1 hr/day on outreach, Fact: 3.5 hrs — Gap: 2.5 hrs on manual CRM updates and lead list building.
• Support lead: Planned 2 hrs/day on ticket triage, Fact: 4 hrs — Gap: 2 hrs on repetitive password resets and status checks.
• Ops lead: Planned 3 hrs/day on scheduling, Fact: 5 hrs — Gap: 2 hrs on spreadsheet juggling and shift change notifications.
• Marketing lead: Planned 1.5 hrs/day on content planning, Fact: 3 hrs — Gap: 1.5 hrs on asset hunting and version tracking.
• Finance lead: Planned 1 hr/day on invoicing, Fact: 2.5 hrs — Gap: 1.5 hrs on data entry from PDFs and manual matching.

The gap isn’t laziness — it’s system friction. AI agents can handle the repetitive parts, freeing people for judgment, creativity, and customer contact.

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

A founder of a 60-person logistics company ran a 7-day diagnostic on their dispatch team. They expected to find time lost to route planning — but the Fact showed 30% of each day was spent manually checking driver locations via phone and updating customers via WhatsApp. The Gap was clear: a simple AI agent for location check-ins and automated ETA updates could reclaim 10–12 hrs/week per dispatcher. Within two weeks, the team had built and tested the agent using plain language — no code. The founder stopped guessing where delays came from and started seeing real-time updates in their daily digest. Decisions about rerouting or customer alerts shifted from reactive to proactive — all without adding meetings or micromanaging.

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

What if my team resists logging their work? Frame it as reducing their overload, not monitoring them. Keep it light: one sentence per task, no forms. Most people engage when they see it leads to less repetitive work.

Do I need to do this for every team at once? No. Start with one high-friction area — like support, sales ops, or finance — where routine work is visible and painful. Learn, then expand.

Can’t I just ask my managers what the team does? You can, but you’ll get the Plan, not the Fact. Managers often describe ideal workflows, not what happens when systems break or priorities shift.

How is this different from an audit or a consultant’s process? It’s lighter, faster, and founder-led. No decks, no interviews, no external fees. You get real data in days, not months.

What if I already use a project management tool? Those tools show task completion — not the hidden work: context switching, manual fixes, or workarounds. The diagnostic captures what the tools miss.

Conclusion

Before implementing AI, see what your team actually does. The gap between Plan and Fact isn’t a failure — it’s your roadmap. Start small, learn fast, and let the data guide where AI helps — and where it doesn’t.

If you want a system that surfaces the Plan → Fact → Gap automatically — every day, across the company — see how the 7-day diagnostic works.