Back to Blog
AI Strategy Operations Productivity

I Was the Bottleneck: How a Task Audit Showed Me What AI Could Actually Do

Mathieu Arsenault · 28-03-2026

64% of my work was automatable. I had no idea. A structured task audit across 9 business areas and 55 tasks revealed just how much I was doing that didn't need to be me.

I thought I was being efficient. I had my routines. I knew my workflows. I was getting things done.

Then I did a task audit.

Not a vague "think about what you do" exercise. A structured, row-by-row mapping of every task I performed across every area of my business — marketing, sales, client delivery, operations, admin, team management, data, communications, strategy.

55 tasks. 9 business areas. 64% automatable.

The Problem With "Unnamed Work"

Before the audit, most of my work existed as projects, not tasks. "Build the pitch." "Handle client onboarding." "Do the weekly reporting."

The breakthrough came from decomposing those projects into their actual components. "Build a pitch" isn't one task — it's prospect research, discovery prep, deck building, positioning, scheduling follow-ups, and writing the summary email. Each of those is distinct. Each has a different automation profile.

When everything blurs together into unnamed blobs of work, everything feels manual. You can't automate a blob. You can only automate a named, discrete task.

What the Audit Found

After mapping and scoring all 55 tasks, here's what the breakdown looked like:

Category Count % of Work Examples
Fully automatable 13 24% Follow-ups, content repurposing, metric tracking
Partially automatable 22 40% Proposals, discovery prep, reporting
Not yet automatable 5 9% Video production, complex design
Human-only 15 27% Strategy, relationships, coaching, family time

That means 64% of my tasks were either fully or partially automatable — yet I had automated maybe 5% of them. The opportunity wasn't hidden. It was invisible because it was unnamed.

The Realization

The human-only work — strategy, relationships, creativity, coaching, family time — is the original reason I started the business. It's the work that creates value, builds trust, and moves things forward.

Everything else? It's maintenance. Important, but not me.

I wasn't just losing hours. I was spending my highest-leverage time on work a system could do. That's not a productivity problem — that's a business design problem.

What I Did About It

The methodology is repeatable:

  1. Map every task — name it, don't project-ify it
  2. Score each one — fully automatable, partial, not yet, human-only
  3. Find the quick wins — high-frequency fully-automatable tasks first
  4. Build and test one automation — don't batch, don't rush
  5. Repeat — the next audit reveals the next layer

Try It Yourself

Pick one area of your business. List every task you performed in the last two weeks in that area. Then ask for each one: "Does this actually need to be me?"

You'll be surprised what you find.