Free Resource

The Work Audit

A one-week diagnostic that tells you two things: what work to protect, and where AI can actually help. The same audit I run with consulting clients, simplified into a tracker you can fill out in an afternoon.

5 days to run, 15 min to score Excel tracker + rubric + 8 prompts Filled example included

Free. No credit card. Email required once — after that, all downloads unlock.

The core idea

You cannot delegate work you don't understand. The audit gives you that understanding.

This is not about finding things to automate. It's about developing the self-knowledge to make good decisions about where AI creates value, and where it doesn't. Some of the most important outcomes are the things you decide NOT to hand to AI.

PROTECT

The hours in your week that only you can do. Judgment, relationships, accountability. Block them on your calendar before anything else.

LEVERAGE

The hours where AI can help. Fully or partially AI-capable tasks mapped to the right tier of the AI Engagement Spectrum.

1

How to run the audit

One week of logging. One 15-minute scoring session. Two numbers that change how you think about your work.

  1. Day 1-5

    Log every task that took more than 10 minutes

    Don't filter. Don't judge. If you spent the time, log it. Batch at 5pm or use voice memos during the day.

  2. Friday

    Score each row on three dimensions

    AI-capable (Fully / Partial / No), Human Requirement (why does this need a human?), Tier (which AI Engagement Spectrum tier fits).

  3. Review

    Read the Summary tab

    Your PROTECT and LEVERAGE numbers calculate automatically. Tier distribution and human-requirement distribution show where to invest first.

  4. Monday

    Block your PROTECT hours. Automate one Fully task.

    One task, not ten. Let it compound. Re-run the audit in 90 days.

2

The AI Engagement Spectrum

Every AI-capable task lives somewhere on this spectrum. The tracker assigns each task to a tier so you know what to build, not just what's possible.

T1

Assistant — you drive, AI responds

Fast answers, drafts, summaries. You evaluate every output.

T2

Agent — multi-step work toward a defined goal

You write the brief, the agent executes, you review the result.

T3

Coworking — AI as thinking partner that knows your context

Ongoing collaboration, iterative refinement, AI holds state.

T4

Embedded — AI inside a tool your team already uses

Features in M365, Salesforce, Notion, etc. No new tool to learn.

T5

Workflow — runs on a trigger, no human present

Recurring processes with clear inputs and measurable outputs.

T6

Custom Build — proprietary data or domain-specific training

Build only if T1 through T5 can't solve it.

Keep Human

Judgment, relationships, accountability

AI may assist at the edges, but you own this. Protect the hours.

3

Human requirement categories

For every task, tag the dominant reason a human is needed. Tasks tagged Execution or Communication are strong AI candidates. Tasks tagged Judgment, Relationships, or Accountability almost always stay human.

Judgment

Context, values, priorities that aren't in any document.

Relationships

Trust, rapport, political awareness.

Accountability

Someone has to be responsible, and that someone is you.

Creativity

Original thinking, not pattern-matching.

Domain

Specialized knowledge AI has in general but not for your context.

Execution

Clear inputs, defined steps, predictable output.

Communication

Moving information from one place to another in the right format.

4

What's in the kit

The Work Audit — Blank Template.xlsx

The tracker. Four tabs: Start Here (intro + spectrum reference), Tracker (100-row task log with dropdowns), Summary (auto-calculated PROTECT and LEVERAGE), Value Inventory (capability statements).

The Work Audit — Filled Example.xlsx

A real audit week. 43 tasks, 25.4 hours logged, all scored. Reference when yours feels off.

Scoring Rubric.md

The 30-second test, edge cases, common scoring mistakes, and rules for your Friday scoring session.

Scoring Prompts.md

Eight AI prompts that speed up scoring: Batch Scorer, Protect Pass, Tier Translator, Pattern Finder, Partial Upgrader, Honest Audit, 90-Day Replay, Category Drill-Down.

How to read your numbers

Once you've scored the week, two numbers come out of the Summary tab. Here's how to read each one.

Your PROTECT hours

Hours per week in Keep Human tasks. Work that depends on your judgment, your relationships, or your accountability.

Under 5 hrs You're either scoring too aggressively toward AI-capable, or your role really is execution-heavy. Run the Honest Audit prompt.
5 to 10 hrs Typical for individual contributors with clear scope. Defend these hours.
10 to 20 hrs Typical for managers, consultants, senior roles. This is where your value lives.
Over 20 hrs Either you're in a highly relational role, or you're over-scoring Keep Human. Re-check the rubric.

Your LEVERAGE percentage

Percentage of tasks that are Fully or Partial AI-capable.

20 to 35% You're early. Big runway. Start with one Fully task this week.
36 to 50% You've got some leverage, but it's ad hoc. Formalize 2-3 workflows.
51 to 65% You're running hot. Partial is where the next 90 days of growth lives.
66 to 80% You've built real leverage. Audit the Keep Human column. Some of those are probably liars.
81%+ Either you're running an AI-heavy function, or you're over-scoring. Re-run with stricter lines.

Need Help Running This Across A Team?

The individual audit shows one person's work. The team audit aggregates across a group and finds the AI investments that actually matter. If you want structured help running an audit across your team and building a quarterly AI roadmap from it, let's talk.