What is an AI-generated resume, and why does it matter?
An AI-generated resume is a CV written or polished with a chatbot, anything from a full draft to a few rewritten lines. It matters because polish makes every CV read well. The wording stops telling candidates apart, and a claim can grow on the way through.
A polished CV is the first thing a candidate sends you. How candidates use AI later, in the interview itself, is in the guide to interview cheating.
AI help changes some parts of a CV far more than others:
| Part of the CV | What AI help does to it | What still tells you something |
|---|---|---|
| Wording | Smooth, generic, and alike from one CV to the next | Nothing much; stop reading the prose for signal |
| Skills list | Mirrors the words in the job description | Which skills turn up again in the projects, and in the interview |
| Achievements | Rounder numbers and bigger verbs ("led", "drove") | The specifics behind them: what, how big, what went wrong |
| The candidate's own account | Unchanged: a chatbot can't take the follow-up question for them in a live round | The answer when you ask about one claim |
Source: this guide's reading, 17 Nov 2026
Using AI to write a CV isn't lying. Plenty of recruiters use the same tools for job descriptions and outreach. The risk is narrower: a claim that got bigger in the rewrite, such as a team member who now "led" the project, or two years with a tool that became five.
Can you spot an AI-written CV?
Not reliably. The usual tells, such as generic verbs, round numbers and a structure that copies the job description, show polish rather than who wrote the CV, and a candidate who writes that way leaves the same marks. AI-text checkers give a likelihood, not a verdict, so treat a suspicion as a reason to ask about a claim.
Signs worth a question
- An achievement with a number but no context around it.
- A skill in the list that appears in no project.
- Lines that repeat the job description word for word.
- A claim much bigger than the years behind it.
- Two CVs with near-identical lines, which happens in agency-sourced batches.
Each of these is a reason to ask about that claim. None is a reason to reject.
AI-text checkers
Tools that judge whether text was written by AI give you a likelihood. No vendor we read for this guide publishes how often its checker is right, with a method a buyer could check, so this guide quotes no figure.
How do you check the claims on an AI-written CV?
Test the claims, not the prose. Pick the three or four claims the role rests on, ask the candidate for the specifics in the interview, follow up once, and write down whether each answer backs the claim up, overstates it or contradicts it. That works the same whether a person or a chatbot wrote the CV.
- Pick the claims the role rests on: a skill, a scale, years with a tool.
- Ask for one specific: what they built, how big it was, what broke, what they'd change.
- Follow up once on the detail they give.
- Note each claim as backed up, overstated or contradicted, with the answer beside it.
- Score the answers against your scorecard, not against how polished the CV is.
- For facts that talk can't show, such as an employer, dates or a degree, check references or documents (see the next section).
An illustration, invented for this guide: a CV says "led migration to Kubernetes". Ask what the cluster ran, who else worked on it and what failed in the first week. An answer that describes deploying to a cluster someone else ran overstates "led". It doesn't make the candidate a poor hire; it tells you what to weigh.
Questions to ask any screening vendor about AI-written CVs
Take these to any tool you are weighing, Recagn included:
- Does it claim to tell AI-written CVs apart? On what test, and published where?
- What does it compare a CV's claims with: other CVs, a database, or the candidate's own answers?
- Does a finding change the score, or sit beside it?
- Can a finding reject a candidate on its own?
- Does your client see the findings?
- What does it add to the price of a screen or an interview?
When is an interview the wrong check for a CV claim?
When the claim is a fact the candidate can't show by talking: an employer, dates, a degree or a licence. Check those with references, the issuing body or a background check. Use the interview for what someone can explain, such as a skill, a scale or a decision they made.
- Employers and dates. A reference call to a previous manager, or a background-check provider, settles these. An interview answer can't.
- Degrees and licences. The issuer's records are the check. In India, DigiLocker, a platform for storing, sharing and verifying documents and certificates, is one route where the issuer uses it.
- Roles where writing is the job. A CV polished by AI says nothing about how the candidate writes. Set a short writing task in a live round.
- Someone else in the interview. If the question is who is answering rather than what the CV says, see the guide to proxy interviews.
How Recagn does this
Recagn is our product. Its CV claim check, which the product calls the Consistency Checker, sets each claim on the CV against the candidate's interview answers, criterion by criterion, once the interview is complete. Each finding is a corroboration, an inflation or a contradiction, and quotes both the claim and the answer.
Before the interview, Recagn reads each CV against the rubric and treats every line on it as a claim to test, not a fact; that read is part of Recagn's resume screening. The claim check then takes the skills, years, projects and education on the parsed CV and compares them with the job description, the rubric and the candidate's answers, follow-ups included. It reads claims, whoever or whatever wrote the CV.
Read what the CV says, then what he said when asked. Recagn counts what it found; the counts never change his score.
Vikram S. Senior Java full stack and platform engineer
From the CV we wrote
Designed the company's Terraform estate across AWS; 7+ years of Terraform in production.
His answer · Describe a time you designed or optimised AWS infrastructure using Terraform or Ansible. What did you optimise for (cost, availability, security) and what was the measurable impact?
I own our whole AWS Terraform estate and have done this for over seven years. I optimise for cost and availability. We use reserved instances and autoscaling, and everything is in modules. When I joined in 2019 we were still creating most things from the console, and Terraform came in later, but I led that move. The impact was significant savings across the board.
From the CV we wrote
Authored the internal developer portal on Backstage, including 15 custom plugins.
His answer · Tell us about your experience administering Kubernetes or OpenShift clusters and, if applicable, building internal developer portal features with Backstage. What problems were you solving?
I run a platform of twelve clusters on Kubernetes and OpenShift serving hundreds of microservices. I built our developer portal on Backstage with fifteen custom plugins. The main problem we solved was giving teams one place to see everything about their services. The plugins integrate with our tools across the stack.
Profile against interview: 1 contradiction, 4 inflations, 7 corroborations No effect on scores.
Your decision: not made yet
"7+ years of Terraform in production" and "15 custom plugins" are the kind of claim polish makes bigger, and the kind an interview can test: the question that tested each one sits beside it. Vikram S.'s CV was written for the example. The counts cover all his answers, not only the two shown.
The same run, all four candidates:
| Candidate | Recagn's recommendation | Contradictions | Inflations | Corroborations |
|---|---|---|---|---|
| Arjun M. | Disqualified | 0 | 2 | 4 |
| Rohan D. | Qualified | 0 | 2 | 10 |
| Sneha K. | Disqualified | 2 | 4 | 5 |
| Vikram S. | Disqualified | 1 | 4 | 7 |
Source: Recagn's senior Java full stack worked example, run 8 Nov 2026; invented candidates, CVs written for the example
Every candidate had claims the answers backed up and claims they overstated. The recommendation came from the scored answers; the findings sat beside it.
In the interview itself, the Authenticity Detector judges each answer for LLM-written text, a separate signal from the claim check; see Recagn's AI proctoring.
You read the findings in two places. On the interview, a "Consistency" fold lists each one: the criterion, the line quoted as Résumé: "…", the Answer: "…" and the reasoning. On the candidate's record, "Consistency history" keeps them, with the hint "Profile against interview. No effect on scores." The same counts appear on the Company Report you share with your client.
A finding changes neither the interview score nor Recagn's recommendation, and it never rejects anyone. On a later role that tests the same skill, findings can move the candidate up or down the list. Testing every claim against the answers comes off your plate. Which inflation matters for this client stays yours. Recagn recommends. Only you can reject a candidate.
The claim check runs on the interview's credits, so it costs nothing beyond the interview. How it fits into screening is on Recagn's resume screening page.
Questions people ask
Should you reject a candidate who used ChatGPT on their CV?
No. Using a chatbot to write a CV isn't lying, and the wording tells you little either way. Judge the claims instead: ask about them in the interview, and decide on the answers.
Can a recruiting tool like Gem tell if a resume was written by AI?
Gem says not reliably. Its own guide to AI-generated resumes (posted 21 Apr, year not shown; read 17 Nov 2026) calls the accuracy of AI-text detection services "questionable at best" and advises against rejecting a suspected AI CV automatically. It suggests asking detailed questions about specific claims in the interview instead (Gem).
What does a CV claim check cost?
On Recagn, nothing beyond the interview. The claim check runs on the credits the interview already uses, at most ₹50 for a text interview on Starter before GST. How it fits into screening is on Recagn's resume screening page.
Do AI-written CVs break keyword screening?
Partly. A chatbot can match a CV's words to the job description in seconds, so a keyword match now tells you less. Read each CV against the role's criteria, and let the interview test the ones that matter.