The Working Picture
This isn't a test. There's no score at the end, no traffic-light rating, no percentage that tells you how worried to be. It's a way of seeing your working life more clearly than your job title allows — and that clearer picture is what makes any next decision a better one.
It works in four steps, each building on the last. You can stop after any of them. Doing two out of four still leaves you with something useful.
You own everything you write here. Nothing is sent to us. Your responses stay on your device. You can come back and change them. You can copy them and take them wherever you need them.
Take your time. Everything else in the AI conversation is telling you to act now. This is telling you to see clearly first.
Step 1: What You Actually Do
Most of us describe our work by job title. "I'm a project manager." "I work in compliance." "I run the marketing." But that's the label, not the work. This step asks you to look at what actually fills your week.
Think about your last full working week. Not an ideal week — an actual one.
These are the visible tasks — the things that would appear in a handover note. Answering emails, writing reports, updating spreadsheets, attending meetings, processing applications.
The work that doesn't have a tidy name. Reading the mood in a meeting and knowing when to push back. Figuring out that a client's real concern isn't the one they stated. Smoothing something over before it became a problem. Knowing which of three processes to follow when none of them quite fit.
Not a trick question — sometimes the slow, low-value work is genuinely necessary. But sometimes it's just the way things have always been done, and nobody's questioned it.
Not because of your job title, but because of what you specifically know, who you know, or how you read the situation. The thing where someone else could have followed the same process but wouldn't have got the same result.
Take a few minutes with each of these. Write freely — this is raw material, not a presentation. The list you build here is what you'll work with in the next step.
Step 2: How Exposed Is Each Part?
Now take what you've written in Step 1 and hold each activity against the question: how much of this could an AI system do, realistically, in the next few years?
Not in theory. Not in a vendor demonstration. In the actual messy reality of your workplace, with the data you have and the people you work with.
There are roughly four bands. Most people's work sits across several of them, and the boundaries blur.
The routine tasks
Structured, repeatable, rule-based. Data entry. Formatting reports. Scheduling. Standard correspondence. Invoice processing. Basic research — pulling together information from known sources.
These are the most exposed. Many are already being automated. If a task follows a clear pattern and the inputs are predictable, AI handles it well.
The trained judgement
Work that follows established frameworks but needs professional assessment. Routine legal review. Standard financial analysis. First-pass compliance checking. Producing reports from templates where the template is fixed but the content requires knowing what matters.
Highly exposed and accelerating. AI is getting better at this faster than most people expected. But "better" doesn't yet mean "reliable without checking."
The situational calls
Work that requires reading context, navigating ambiguity, or making decisions where the framework doesn't quite apply. Knowing when a client relationship is going south before anyone's said anything. Managing competing demands from people who each think theirs is the priority. Knowing when to escalate and when to handle it yourself. The judgment that comes from experience rather than training.
Less exposed. AI struggles here — not because it can't process the information, but because these calls depend on understanding things that aren't in the data: organisational politics, personal dynamics, the difference between what people say and what they mean.
The knowledge only you carry
Built over years, embedded in relationships and understanding. Trust networks — the people who pick up the phone because it's you. Institutional memory — knowing why the current process exists and what happened when someone tried to change it. Cultural navigation — understanding how things actually work versus how the org chart says they work.
Least exposed. Genuinely hard to replicate. Not because AI isn't clever enough, but because this kind of knowledge isn't written down anywhere for it to learn from.
Complete Step 1 first — your activities will appear here for mapping.
Go through your list from Step 1. For each activity, ask yourself which band it sits closest to. Don't agonise over the edges — many things will span two bands. The point is the overall picture. How much of your week sits in the more exposed bands? How much in the less exposed ones? And — crucially — where does the work that matters most sit?
Step 3: What Your Job Title Doesn't Show
This is the step most people find surprising.
Organisations shape how you see yourself. The job title tells you what you are. The annual review tells you how well you're doing it. For many people, that becomes the only way they understand their own professional worth. But the role isn't the person. You have knowledge, capabilities, and judgment that your current position may never have asked you to use or name.
These questions are designed to surface those things. Sit with them. There's no rush.
Not just colleagues — clients, contacts, people from other teams. What do they trust you with that goes beyond your formal remit?
Not the tasks that would go unassigned — the things that would stop working smoothly because you're the one who understands why they work in the first place.
The skills from a career before this one. The things you're good at that your current role never calls on.
Experience from elsewhere. Knowledge from a different field. An instinct that came from somewhere other than your training.
Write as much or as little as you want against each. Some of these may take a while to answer honestly. That's the point — these are the things that have been invisible, and making them visible changes the picture significantly.
Step 4: Questions to Sit With
You've built a picture across the first three steps. Before you do anything with it, sit with it.
There is no urgency here that can't wait for a clear view. Everything else in the AI conversation is telling you to act now. This is telling you to understand first.
These questions don't have right answers. Some will open doors. Some will confirm what you already suspected. Some may be uncomfortable. All of them are more useful than the question the headlines are asking, which is "should I be worried?" — because they replace a feeling with a picture, and a picture is something you can act from.
What to do next
The Working Picture is yours. Take it to a conversation with someone you trust — a colleague, a mentor, a friend who knows your work. Or sit with it alone and see what it tells you over the next week. If you want to go further, the book The Next Move builds the full framework around this exercise. The skills section of this site offers structured ways to make the capabilities you've surfaced here visible, documentable, and transferable.
If you run a business, the Business Reality Audit looks at your firm through five practical lenses.