Someone Is Going to Automate Your Job — Make Sure It's You

The question most people are asking about AI is: will it replace my job? The more useful question is: which parts of my job should I be replacing — and am I the one doing it?

Those are very different questions. The first is something that happens to you. The second is something you drive. And right now, the gap between people who are driving it and people waiting to see what happens is opening fast.

The spectrum

The honest picture is a spectrum. At one end: repetitive, rules-based work with known right answers. A CSM manually answering the same onboarding question for the hundredth time. An SDR spending half their week on research and CRM updates. A support team rewriting the same reply to the same email. That stuff should be automated.

At the other end: judgment, creativity, context, nuance. A designer deciding what’s actually right for the user. An engineer making an architectural call. A salesperson reading a room and adjusting mid-conversation. Work that requires someone who has been doing this long enough to know where the edges are. AI can assist here, sometimes significantly, but it’s not replacing the judgment. Not yet.

Most jobs contain both. What matters is whether you know where each part of your work sits — and whether you’re doing anything about the parts that should be gone.

This looks different depending on your role. If most of your work is repetitive and execution-heavy, the opportunity and the urgency are obvious. If most of your work is judgment and creativity, the shift is subtler — AI becomes a thinking partner that compresses the execution parts rather than a system that takes over. Either way, the question is the same. You just have different answers.

Experience is what makes this work

The people best placed to automate the right parts of their job are the ones who’ve done it long enough to know which parts those are.

You can’t judge AI output on something you’ve never done deeply. If you’ve onboarded a hundred customers, you know immediately when the AI’s answer is missing something crucial — the edge case it didn’t account for, the tone that won’t land with this particular person, the thing that sounds right but isn’t. That judgment is expensive to acquire. It takes time and repetition and getting things wrong. There’s no shortcut to it.

I’ve written about this before — that taste and domain knowledge are what separate people who use AI well from people who just use AI. The practical version: without depth of experience, you’re automating blind. You produce output that looks plausible and miss that it’s wrong. The dangerous part isn’t the mistake — it’s that you won’t see it.

Where the next generation comes from

There’s a tension here that doesn’t get talked about enough.

If AI handles the entry-level execution work: the repetitive tasks, the first drafts, the routine answers — where does the next generation of experienced people come from? Junior lawyers learn by doing the contract work. Junior CSMs learn by doing the onboarding calls. Junior designers learn by doing the production work. If those tasks disappear, the pipeline of people developing the depth to judge AI output eventually dries up.

Juniors coming into the workforce now are often more technically fluent than anyone who came before them. They may lack domain experience but they’re AI-native in a way most senior people aren’t. That creates an interesting skills gap — experience without fluency on one side, fluency without experience on the other. The most valuable people will be the ones who have both. Right now, that’s rare.

By the time experienced people become truly scarce, AI will likely have moved far enough up the value chain that it’s encroaching on the judgment work too. The window where deep human experience is the primary differentiator may be shorter than anyone is comfortable admitting. Which makes using it now — to rethink how work gets done — more urgent than it might feel.

So where does your work sit?

Worth thinking through with a specific task rather than in the abstract. Not your job as a whole — one thing you do regularly.

Is it repetitive? Does it follow the same steps each time? Is there a known right answer, or is the quality subjective and contextual? How much does hard-won experience in this specific role matter to doing it well? How much creative judgment is involved? The more repetitive, rules-based, and context-light, the less it needs you. The more it requires judgment and earned context, the more it needs you, and the more AI should be a thinking partner rather than a replacement.

The answer to “which parts of my job should disappear?” will keep changing. Staying ahead of it isn’t a one-time audit. It’s an ongoing way of thinking about what you do.

Making it happen — and who gives you the time

I’ve written before about documenting answers so you never have to give the same one twice. That’s one version of this thinking. But removing yourself from a task goes further. It means mapping what you actually do, being honest about which parts need you and which don’t, and then building the workflow, the agent, or the process that handles the rest.

The uncomfortable reality is that most employers won’t give you the time to do this. Not because they don’t want it done — because the day job is always louder than the work of improving the day job. The task in front of you wins. It always does.

So what do you do? You carve it out yourself. Not a big commitment — an hour, one experiment, one task that frustrates you. Most of the genuinely useful workflow improvements I’ve seen come from one person who got tired of doing the same thing the same way and just went and fixed it. Not from a company initiative. Not from a training programme.

The time you spend removing yourself from a task is the highest-value time you can spend on it. Most people never find it because the task is always there, always urgent. That’s the problem — and nobody is going to solve it for you.

If you’re leading a team

The individual contributor question is personal — which parts of my job should disappear? Leaders have a harder version of it: which parts of my team’s work should disappear, and am I creating the conditions for that to happen?

A leader who hasn’t figured this out for themselves can’t effectively lead others through it. You can say all the right things about automation-first thinking, but if your team watches you still doing things the old way, it won’t land. The most powerful thing a leader can do right now is model the behaviour, not just mandate it.

Protect time for experimentation, not just add it to a list of things people should do on top of everything else. Recognise when someone ships a workflow improvement as something worth celebrating — more than just shipping more of the same work.

The harder question is this. If your team adopts this thinking seriously, the nature of the work changes. Some roles evolve. Some people will adapt and become more valuable. Others won’t — not because they’re not good, but because adapting requires a kind of curiosity and drive that can’t be mandated. A leader who is genuinely thinking about this will eventually have to reckon with the shape of the team they need versus the one they have. That’s uncomfortable. It’s also unavoidable.

The people who don’t

The people who keep doing their job exactly as they’ve always done it, not because the work requires it but because they haven’t stopped to ask whether it should, are going to find themselves in a difficult position. Not immediately. The gap compounds slowly. Then it’s obvious.

Someone who has automated the repetitive parts of their role is now doing more of the high-judgment work — more of it, better, faster. They’re not more talented. They just asked the question earlier and did something about it. That person is more valuable now. The person next to them doing things the old way isn’t worse. They’re just increasingly less of what’s needed.

Companies aren’t trying to replace people. They’re trying to win. The people who help them win by rethinking how work gets done are the ones who stick. The ones who don’t will find the argument for keeping them gets harder to make. The numbers tell their own story.

Someone is going to automate your job. The only real question is whether you get there first.