AI can write a flawless CV in seconds, so every candidate now looks like a star. Here's how to find the ones who genuinely are.

Amy Wilson-Wyles

For most founders and CEOs, hiring isn’t a project you finish. It is a background process that never stops running, and is often quoted as one of the key topics that keeps leaders up at night. Get a hire right and they compound: they raise the bar, pull in other good people, and buy you time to think about the business instead of babysitting it. Get one wrong and the cost is brutal, and rarely just financial. You lose months of momentum, a team quietly starts to check out, and you now have to dig your way back out while everything else keeps moving. At an early or scaling company, a single bad hire can genuinely bend the trajectory of the whole business.

Which is what makes this particular moment so uncomfortable. The CV, the thing most of us still reach for as a first filter, has in many ways stopped doing its job. Anyone can now generate a polished, keyword-perfect CV in seconds, so inboxes fill with immaculate, near-identical applications that all read like the ideal candidate. When everyone looks exceptional on paper, the paper tells you almost nothing. So how do you find the people who genuinely are?

We put that to three firms who hire talent for a living. Forme Partners is a London executive search firm that works exclusively on C-suite, revenue-driving roles for high-growth tech companies, partnering with founders and investors. La Fosse has been one of the UK's best-known technology, digital and change recruiters since 2007, hiring at every level from startups to the FTSE 100. Propel is a London talent partner that has spent two decades helping venture and private-equity-backed tech businesses build their commercial, product and technology teams. Here is what they told us, and what you can actually do with it.

Stop trying to spot the AI. It is a losing game

The instinct is to hunt for tells. Everyone knows them by now: the verbs that travel in packs like "pivotal", "quietly", "spearheaded" and "orchestrated," the way whole parts of the job spec are magically woven into their CV. But Kate Zatland, Founder and Managing Partner of Forme Partners is blunt about where that leads. "Those obvious AI signals won't be around for long. The tools get better every month, and a well-built CV won't have any tells left."

Tim Clark, Associate Director at La Fosse, where he leads the Data, AI & Engineering practice, thinks that attempting to play the detection game is already futile. AI, Clark says, "has raised the baseline quality of CVs across the board," helping candidates structure their experience, tighten their language and present themselves more professionally than many could manage alone. "That's not a bad thing." Ben Dalzell, COO of Propel, agrees, and would actively encourage candidates to use it, "but only if they know how to use it well," as a tool for structure and clarity rather than a substitute for thinking.

The practical takeaway is simple: do not waste your screening time playing forensic linguist. Assume every strong CV has had AI help, and change what you are looking for instead.

Look for specificity, not style

If the writing no longer separates candidates, what does? For Clark, it is authenticity. AI-written CVs read well but tend to share a shape: polished, generic in places, light on real detail and short on measurable impact. "You'll see a lot of confident language about what someone 'drove' or 'delivered', but very little about how they actually did it, what the constraints were, or what they'd have done differently. The tell isn't the writing style. It's the absence of specificity."

So what you are reading for is texture: an unusual project, a slightly unexpected number, a detail that only makes sense if the person was actually in the room. Dalzell has a blunter name for the opposite of that. "The biggest giveaway is waffle," he says. "If a CV is full of generic statements, overcomplicated language and vague achievements, it immediately raises question marks." Clark agrees and suggests looking at every CV through this lens: if a CV could belong to anyone in that role at any company, that is the real red flag, not any single word choice or formatting quirk.

Move fast from the document to the conversation

If a CV now only tells you who can write a good CV, the job of finding talent moves into the interview, and the experts are specific about how to run it. Clark's approach is to get off the document quickly and into real scenarios that are hard to fake: "walk me through a decision you got wrong," "tell me how you'd approach this specific problem," "show me your working." The point is to test what someone actually did and how they think, rather than how well their achievements have been packaged.

The candidates who hold up, he says, "back up their claims with specific examples, measurable outcomes and a clear account of the role they personally played, not the team, not the project, them." Just as tellingly, they can describe what did not work, "which is often more revealing than what did." Dalzell frames the same instinct from the candidate's side. "The strongest CVs aren't necessarily the best written," he says. "They're the ones that make it easy to understand the value someone has brought to a business, and showcase the commercial results."

The two things AI can’t fabricate

For senior hires especially, Zatland argues that the strongest evidence sits entirely off the page. As the firm puts it, "there are two things AI can't fabricate."

The first is tenure, or in Zatland's words, "the actual shape of someone's career over time. Did they stay long enough to live with the decisions they made? Or did they leave right before the results came in?" A CV can list a run of impressive titles. Only the timeline tells you whether the person owned their outcomes or moved on before the bill arrived.

The second is what a candidate's network says about them, and this is where reference calls earn their keep, provided you ask the right questions. Zatland points our some key ones: "How they respond when met with resistance from their board; whether their old team would take a pay cut to work for them again; or what they actually did when things didn't go to plan." The value is in the unvarnished account. "A former boss, a peer or someone who watched them under real pressure can give you the unfiltered version of what to expect."

"A standout candidate isn't the one with the most polished CV. It's the one whose tenure and network tell the same story off-paper."

The CV was never the point

The reassuring part, for anyone staring at an inbox of flawless applications, is that none of this asks you to out-clever the AI. It asks you to demote the CV back to what it always really was. As Clark puts it: "The CV gets you the interview. It's never been what gets you the job, and that's more true now than ever." AI has raised the noise. The evidence, specificity, the shape of a career and the verdict of people who have seen someone under real pressure, is exactly where the best hirers were always looking.

What this means for your next hire

• Assume AI touched every CV, and screen for substance over style. Read for the "how," the constraints and the person's individual role, not for confident verbs.
• Treat vagueness as the red flag. If a CV could belong to anyone in that role at any company, it is telling you nothing.
• Get off the document fast in interviews. Ask for a decision they got wrong, their live working on a real problem, and the reasoning behind an actual call they made.
• Make references do real work. Ask whether the old team would follow them, how they handled pushback from the board, and what they did when things went sideways.
• Read the timeline. Did they stay long enough to own the outcome, or leave just before the results were really felt?

‍