Originally published on LinkedIn in December 2025. Republished here for completeness.

We’ve all been there. It’s that time again where you head out on the job search and need to update your CV. You open the last one, click File -> Save As, then go down to the summary and start shoving in the latest brain dump of your most recent role.

STOP RIGHT THERE!

Last time I talked about not having a toaster give a TED Talk on your behalf. Today I’m going to show you how to use AI, or more precisely, an LLM, to help you dig out some extra gems to add to your CV. Now we make the toaster do some work.

AI is a tool, but it NEEDS to be told EXACTLY what we want, how we want it, and what we don’t want (such as buzzwords and empty cliches).

Also, AI isn’t you, and while we can make it try and sound somewhat like you, it won’t. This means you need to take some ownership, do some editing, and make it yours.

Enough preaching. It’s time to do some work.

One of the newer things that AI is being used for is as a guided interviewer, which is then going to give you some feedback. It’s already being used for career coaching, finding your values, and even helping to resolve problems.

Today, you’re going to get the AI to read in your existing CV or resume, and for it to ask iterative questions for each role. The aim is to assess what was done, how the impact was measured, why it mattered, and what tools, techniques and approaches were used to achieve it.

Once it’s gone through that, you’re going to ask it to give you bullet points or accomplishments that should get picked up by the ubiquitous ATS, and make hiring managers race to call you.

This is going to take a couple of steps.

  • upload the voice profile you created last time, and
  • tell the AI that you’re going to upload your CV and what you want done with it.

First up, let’s upload your voice profile - your personality chip. Copy in the following prompt, then paste in your voice profile after that.

This is my voice profile. For all future responses in this conversation, use this voice and style unless I explicitly tell you otherwise. Do not rewrite anything yet — just confirm you understand my voice profile by summarizing it in one sentence. The model should reply with something like “Understood. I will write in this style …” and summarise the tone in one sentence.

Now it’s ready to actually be useful, and this is where the fun starts. If your career is as long as mine – it might pay to tell it to limit the focus, or stop it at any time. And because you told it to use your style for the rest of the conversation – it’ll ask questions like you would.

Cut and paste the following prompt into the AI.

I’m going to upload my current CV/resume. Your job is to act as a skilled career interviewer. Don’t rewrite anything yet — start by asking ONE question at a time. For each role, help me extract clearer, stronger achievements by asking about: my key responsibilities; my most important achievements; specifics for each achievement (problem, what I did, tools used, measurable improvements, and why it mattered); skills I developed; scope of my work (team size, data volumes, budgets, users, etc.). Keep questioning until each achievement is specific and measurable. When I confirm I’m done, summarise everything into achievement bullet points using: verb + what I did + how I did it + measurable result + business impact. Then upload your CV.

It should think for a moment while it digests what’s already there, then start asking questions. Once you’ve finished, or run out of time or caffeine, tell it you’re done and it should drop out a series of bullet points for each role.

You WILL need to edit this. Nobody expects you to be as perfect as the AI thinks you are. But it should be close. If you’ve followed along, you’ve now seen how an LLM can do more than spit out generic “dynamic professional with excellent communication skills” nonsense. Used properly, it becomes a slow-burn interviewer that drags the specifics out of you: what you actually did, how you did it, why it mattered, and where you made a measurable dent.

The voice profile keeps the AI sounding like a human (specifically, you), and the structured questioning keeps it honest. The end result isn’t a magically “perfect” CV — it’s a sharper, clearer, more accurate version of your work history that you still need to polish with your own judgement. AI gives you the raw material. You finish the job.

If your CV used to be a timeline of job titles, this process turns it into something that actually says what you’re capable of.

Next time, I’m going to switch gears and show you how to use AI to figure out what your CV is missing.