There is a specific moment in interviews that has become uncomfortably common.
The interviewer asks a simple follow-up about something on the CV. Not a trick question. Something like: you mention leading the CRM migration — what was the hardest call you had to make on that?
And the candidate's face changes. Because they did not write that line. A model did, from a job description, and they have not thought about it since.
This now happens several times a week in busy hiring processes. It is the defining feature of hiring in 2026, and it has quietly changed what a CV is even for.
The signal collapse
A major 2026 analysis built on interviews with more than a hundred talent acquisition leaders, alongside a review of several thousand recorded screening sessions, reached a blunt conclusion: generative AI has undermined the reliability of the signals hiring depended on for decades. The ability to present a polished CV and deliver confident, structured interview answers used to be a genuine differentiator. It is now essentially free and infinitely scalable.
The consequences appear throughout the data:
- 65% of hiring managers say candidate skills are harder to verify in AI-optimised résumés.
- Around 40% of technology candidates are believed to have meaningfully inflated their résumés.
- Major consulting research in 2026 warned that verification is becoming a core recruiting skill in its own right — as fundamental as sourcing.
- Screening tools bought to manage the volume have their own reliability problem. Industry testing found only 14% overlap in shortlists when the same AI screening tool was run twice on identical candidate data.
That is bad news if your CV is a list of competencies. It is very good news if you can be specific, because the bar for standing out has rarely been lower.
What the verification interview looks like now
Screening has changed shape in about eighteen months. Where interviewers once asked candidates to walk through their experience, the pattern now is to pick one line from the CV and go three questions deep.
1. "Tell me about this project." Anyone can answer this. AI-written CVs survive here.
2. "What was your specific role — what did you personally own?" The first tier falls away. People who were adjacent to a project rather than responsible for it start hedging into "we."
3. "What went wrong, and what did you decide to do about it?" This is the real question. Nobody's CV mentions the thing that went wrong. Only someone who was actually there can answer it, and the answer is always specific, slightly messy, and convincing.
Call it the depth test. Real experience has texture — constraints, personalities, a bad week in March, a compromise you are still faintly annoyed about. Generated experience is smooth all the way down.
You cannot fake texture. You can only have been there.
Why good candidates fail the depth test
Here is the genuinely unfair part.
A large share of the candidates who fail this test actually did the work. They are not frauds. They have simply forgotten the details.
Consider it honestly. Could you state, right now, the specific numbers from a project you finished two years ago? What the baseline was before you changed it? Who objected, and why? What you would do differently?
Most people cannot. The work happened, the evidence evaporated, and by the time they are job hunting they reach for AI to reconstruct a plausible version — not to deceive, but because the real version is no longer accessible to them.
That is the actual crisis. Not fraud. Evidence decay.
Building an evidence trail while it still exists
The solution is unglamorous and it is the highest-leverage career habit available: capture your wins as they happen, not when you need them.
This is exactly what the LeveliU Achievement Vault is for, and of all seven tools in the app it is the one worth recommending most to people who are *not* currently job hunting. You record a win while it is fresh — what the situation was, what you did, what resulted — and the AI helps polish the description into something usable. It stays there, timestamped, ready to export for reviews, interviews and promotion conversations.
Five minutes a week. That is the whole commitment.
The compounding effect is significant. After a year you have a real inventory of your own work. When a job ad asks for stakeholder management experience, you are not inventing an example under pressure — you are selecting the best of eleven documented ones. When your manager asks what you delivered this quarter, you have an answer with numbers in it. When someone asks the depth-test question, you have texture, because you wrote it down while the texture still existed.
Three ways to make the entries genuinely useful:
Capture the messy middle, not just the outcome. "Migration delivered on time" is a claim. "Vendor pushed the timeline three weeks, I renegotiated scope with the finance lead and we cut two nice-to-have modules to protect the go-live date" is proof. The second survives three follow-up questions.
Record the baseline. The most common reason a metric collapses in interview is that the candidate remembers the improvement but not the starting point. "Reduced churn by 12%" invites: from what to what, over what period, on what book size? Write down the before, every time.
Note who else was involved. Not to give away credit, but because "I worked with the Ops lead and one analyst" is far more credible than a story in which you apparently did everything alone.
Turning evidence into an application
Once you have the raw material, the rest gets much easier.
Run your CV through the CV Analyzer — it scores 0 to 100 from a recruiter's perspective and flags which claims read as unproven. That flag list is essentially a preview of the questions you will be asked. Go through it line by line and swap generic statements for Vault-backed specifics.
Then, per role, the Job Match Engine tells you which requirements you can genuinely evidence and which you cannot, with a Go / No-Go verdict. This is where the two connect: a match score means far more when the CV behind it is backed by documented achievements rather than optimistic phrasing.
Before the interview itself, the Interview Coach lets you rehearse the depth test on your own material — practising the three-questions-deep version of your stories until you can hold them comfortably.
The competitive advantage hiding in all this
There is a widespread worry that AI has made it harder to stand out. Having read several hundred nearly identical CVs, the opposite case is stronger.
When every application is polished, polish stops being a differentiator. What differentiates is specificity — and specificity is something only you actually have. The candidate who can say "we cut onboarding from eleven days to seven, mostly by killing a duplicate approval step nobody could justify" is instantly, obviously real.
That candidate is rare now. Being rare is the entire objective.
Start writing things down. Your next interview will be decided by what you can remember.
---
FAQ
How do I prove my skills on a CV without exaggerating?
Replace claims with evidence: baseline, action, outcome, scope. "Improved efficiency" is a claim. "Cut invoice processing from 11 to 7 days across a 40-person team by removing a duplicate approval step" is evidence. Modest and specific beats impressive and vague.
What do recruiters ask to verify experience?
Increasingly they go three questions deep on a single CV line: what the project was, what you personally owned, and what went wrong and how you handled it. The third question is where unowned experience becomes obvious.
How do I remember achievements from years ago?
Mostly you cannot reconstruct them accurately, which is why capture has to be ongoing. Logging wins as they happen takes about five minutes a week and gives you a timestamped, exportable record for reviews and interviews.
Are metrics on a CV still credible in 2026?
Metrics with context are. Metrics without context now actively hurt you — reviewers have seen so many suspiciously round invented numbers that a bare percentage reads as a red flag. Always include the baseline and the scope.
What is the best way to prepare for a competency interview?
Build a documented inventory of real examples first, then rehearse going three questions deep on each. Practising on your own material is far more effective than memorising generic structured answers.
---
Your next interview is decided by what you can remember. Start your Achievement Vault free in the LeveliU app.