The Weekend AI Audit: Map Your Own Job Before Someone Else Does

There's a meeting happening right now, somewhere in your company, where someone with a spreadsheet is assessing which roles can be "optimised" with AI. They're going through job descriptions, matching tasks against AI capabilities, and colour-coding a matrix that will determine who stays and who goes.

You should do this exercise yourself before they do it for you. Because when you do it, you get to act on the findings. When they do it, you get to hear the findings in a meeting room with an HR representative present.

i wish someone had told me to do this audit eighteen months before my redundancy. Not because it would have changed the outcome — the restructuring was coming regardless — but because i would have spent those eighteen months very differently. i would have leaned harder into the parts of my job that AI couldn't touch and automated the parts that it could, positioning myself as someone who was ahead of the curve rather than behind it.

This is a weekend exercise. Literally one weekend. Grab a coffee, open a spreadsheet, and map your own job honestly. What you find might be uncomfortable. It will definitely be useful.

The audit framework

You're going to create a simple matrix. Every task you do in your job, rated on two dimensions: how exposed it is to AI, and how valuable it is to the organisation.

This gives you four quadrants:

High AI exposure, low value: Automate immediately. These are the tasks that make your role look expendable. Get rid of them or make them invisible.

High AI exposure, high value: Automate with caution. Use AI to do these faster and better, but keep your human oversight visible. This is where you demonstrate that AI plus you is better than AI alone.

Low AI exposure, low value: These are annoying but safe. Don't spend too much energy here.

Low AI exposure, high value: Double down. These are your career moat. The things that make you irreplaceable. Spend more time on them. Get better at them. Make sure everyone knows you do them.

Step 1: List everything you do

And i mean everything. Not the version on your job description. The actual tasks you perform in a typical month.

Open a spreadsheet. Create columns for: Task, Frequency (daily/weekly/monthly), Hours per week, Description.

Now fill it in. Be granular. Don't write "reporting." Write "compile weekly sales data from three systems," "format data into standard report template," "write analysis summary," "present findings to management team."

The more specific you are, the more accurate your assessment will be. "Project management" is too broad to assess. "Schedule meetings with stakeholders," "track deliverables against timeline," "write status updates," "resolve blockers by coordinating between teams" — that's assessable.

Most people have between 15 and 30 distinct tasks when they break their job down properly. If you have fewer than 10, you're being too general. Go deeper.

Here's a starter list to jog your memory:

• Email management and responses
• Meeting attendance and participation
• Meeting preparation and follow-up
• Report creation and formatting
• Data gathering and compilation
• Data analysis and interpretation
• Document drafting and editing
• Presentation creation
• Client/stakeholder communication
• Team coordination and management
• Problem-solving and decision-making
• Strategic planning
• Relationship building
• Training and mentoring
• Administrative tasks
• Quality checks and reviews

Your list will look different depending on your role. But everyone does more individual tasks than they think they do.

Step 2: Rate AI exposure

For each task, rate how well AI can do it right now (not in theory, not in five years — right now) on a scale of 1 to 5:

1 — AI can't touch this. Requires deep human judgement, relationships, physical presence, or emotional intelligence that AI genuinely cannot replicate. Examples: negotiating a difficult client relationship, mentoring a struggling team member, handling a sensitive HR situation.

2 — AI is marginally helpful. Might save a few minutes but doesn't fundamentally change the task. Examples: scheduling complex multi-party meetings, facilitating a brainstorming session.

3 — AI can do a first draft. AI handles the initial work, but significant human refinement is needed. Examples: writing a proposal, creating a presentation, analysing data for insights.

4 — AI can do most of it. Human involvement is mainly quality control and edge cases. Examples: formatting reports from templates, initial data cleaning, routine correspondence, meeting summaries.

5 — AI can do this end-to-end. Human involvement adds minimal value. Examples: data entry between systems, simple formatting tasks, standard email responses, basic information lookup.

Be honest. The most common error is rating tasks as 1 or 2 when they're actually 3 or 4. Your ego wants to believe everything you do requires deep human insight. It doesn't. Some of what you do is mechanical, and pretending otherwise doesn't protect you — it just leaves you unprepared.

A good test: if you could explain the task to a smart graduate in 10 minutes and they could do it within a day, AI can probably handle a significant chunk of it.

Step 3: Rate organisational value

For each task, rate how much value it creates for the organisation on a scale of 1 to 5:

1 — No direct value. Necessary overhead that doesn't contribute to any strategic goal. Examples: most internal admin, filing, routine reporting that nobody reads.

2 — Low value. Keeps things running but doesn't differentiate or advance. Examples: standard meeting management, basic compliance documentation.

3 — Medium value. Contributes meaningfully to team or departmental goals. Examples: client communications, project tracking, team coordination.

4 — High value. Directly impacts business outcomes, client satisfaction, or strategic objectives. Examples: client relationship management, strategic analysis that informs decisions, talent development.

5 — Critical value. If this stopped, the organisation would immediately feel it. Examples: key client relationships, critical decision-making, unique institutional knowledge.

Again, be honest. If you stopped doing a task and nobody noticed for a month, it's not high value, however much time you spend on it.

Step 4: Plot the matrix

You now have a list of tasks with two scores each. Put them in the matrix.

Quadrant 1: High exposure (4-5), low value (1-2). THE DANGER ZONE. These tasks make your job look automatable. They're the first things consultants point to when recommending headcount reduction. Automate them yourself, immediately. Use the tools available to you and build simple workflows that handle them. The goal is to make these tasks take minimal time so they no longer define your role.

Quadrant 2: High exposure (4-5), high value (4-5). THE AUGMENTATION ZONE. These are tasks where AI can make you dramatically more productive. Don't avoid AI here — embrace it. Use AI to do these tasks better and faster, but keep yourself in the loop as the quality controller and decision-maker. This quadrant is where "AI plus human" significantly outperforms either alone.

Quadrant 3: Low exposure (1-2), low value (1-2). THE BACKGROUND ZONE. Not worth worrying about. These tasks are safe from AI but also not helping your career. Minimise time spent here where possible, but don't stress about them.

Quadrant 4: Low exposure (1-2), high value (4-5). YOUR CAREER MOAT. This is where your focus should go. These are the high-value tasks that AI can't easily replace. Relationship building. Strategic thinking. Creative problem-solving. Complex negotiation. Mentoring and leadership. Spend more time here. Get better at these things. Make them a bigger part of your role.

Step 5: Make your plan

Based on your matrix, you should now have a clear action plan:

Immediate actions (this month):

• Automate your top 3 Quadrant 1 tasks using AI tools
• Start using AI to augment your top 3 Quadrant 2 tasks
• Identify one Quadrant 4 skill to actively develop

Medium-term actions (next 3 months):

• Reduce Quadrant 1 time by at least 50%
• Redirect saved time into Quadrant 4 activities
• Document your AI-augmented results for performance reviews
• Build at least one visible AI project from your Quadrant 2 work

Long-term positioning (next 6-12 months):

• Reshape your role so that Quadrant 4 dominates
• Become the person who trains others on Quadrant 2 augmentation
• Position yourself as the AI person on your team

What the audit usually reveals

i've walked about thirty people through this exercise now, and the results follow a pattern.

Most people discover that 30-40% of their time goes to Quadrant 1 tasks. Things that are highly automatable and not very valuable. This is a wake-up call. It means that a significant chunk of what you do every week is exactly the kind of work that AI restructuring targets first.

Most people also discover that they spend surprisingly little time on Quadrant 4 — the high-value, low-AI-exposure tasks that should define their role. Often less than 20% of their week. This is the real insight: you're spending most of your time on the wrong things.

The people who act on this audit rebalance their time. Over three to six months, they shift from spending 30% on Quadrant 1 to spending 10%, and from 20% on Quadrant 4 to 40%. Their role transforms from one that's vulnerable to AI to one that's enhanced by AI.

The people who do the audit but don't act on it have a very uncomfortable document sitting on their desktop that accurately predicted what the consultants would say six months later.

Do this, this weekend

1. Block out 2-3 hours on Saturday or Sunday morning.
2. Create the spreadsheet with columns: Task, Frequency, Hours/Week, AI Exposure (1-5), Org Value (1-5), Quadrant.
3. List every task. Be thorough and honest.
4. Rate each one.
5. Sort by quadrant.
6. Make your action plan.
7. Start executing on Monday.

This is the most valuable career exercise you can do in 2026. It takes one weekend. It costs nothing. And it tells you exactly where to focus your energy to stay relevant, stay employed, and stay ahead.

Don't wait for someone else to audit your job. They won't be as kind about the findings as you will.