When every application reads the same: finding real signal in senior hiring

Juan Molina ยท

AI has made CVs, cover letters and LinkedIn posts cheap to produce and hard to tell apart, and candidate fraud is rising. For senior tech, product and talent hires, the signal now sits in references, real work, structured interviews and verified identity.

In June 2025, The New York Times reported that applications submitted on LinkedIn had risen more than 45 percent in a year, to an average of 11,000 a minute (The New York Times, as reported by Ars Technica). The Times said generative AI tools were contributing to the flood.

Volume is only half the problem. The applications also look alike. A CV can now repeat every keyword in your job description, and a cover letter can sound as if the candidate has followed your company for years. Neither tells you much about the person.

I have spent more than twenty years in search. My view is that the CV was always a weak signal for senior roles, and AI has made it close to useless. Here is where I think the signal still is when you hire a VP of Engineering, a Head of Product, a QA or program leader, or a head of talent.

What changed

Sameness. Recruiters now see identical answers across candidates. In a 2024 HuffPost piece, a Zapier recruiting manager said a Zapier application question drew "the same use case of a flower shop" from many people. She put the share of applications that appear AI-generated at "easily 25%" (HuffPost, 2024). That is one recruiter's estimate, not a survey.

Public profiles. Posts used to show how someone thinks. Originality AI, a company that sells AI-detection software, estimated that 54 percent of long-form LinkedIn posts were likely AI-generated as of October 2024 (Originality AI, as reported by Tech.co, 2024). In a July 2026 update on 5,000 public posts, it put the figure at 81.2 percent (Originality AI, 2026). Detector estimates are rough. Still, a polished feed now tells you little.

Fraud. Gartner predicts that by 2028, one in four candidate profiles worldwide will be fake. In a 2025 Gartner survey of 3,000 job candidates, 6 percent admitted to interview fraud, either posing as someone else or having someone pose as them (Gartner, July 2025). In January 2025 the US Justice Department announced indictments in a scheme that allegedly used stolen and forged identities to get North Korean nationals remote IT work at at least 64 US companies. A US "laptop farm" hosted company-issued laptops so the companies believed they had hired US-based workers (US DOJ). It involved IT workers, not executives, but it shows what a remote process can miss.

Senior hiring was never an inbound game

Here is one public example of what inbound volume costs. In 2024, Business Insider described a single level-three software engineering role at a global tech company. It drew 3,360 applicants and seven referrals. The recruiter screened 43 of them by phone, 1.28 percent of the total, and one person got an offer, 50 days after the job was posted (Business Insider, 2024). The first cuts, made with applicant tracking software, removed 1,662 applicants outside the listed hiring locations and 739 who needed visa support.

That was an individual contributor role. For a senior leader, my view is that the person you want is usually employed, doing well and not reading job boards. You find them by mapping who has done the job at companies like yours, asking people you trust who they would hire again, and calling them. In my experience, referrals and direct sourcing have always carried most of the weight in senior search. With inbound now full of generated noise, I would lean on them even more. I search in the US, Europe and Latin America, and the method is the same in each.

Where the signal still is

1. A scorecard built on outcomes, written before you meet anyone

Before sourcing, write down what this leader must achieve in the first 12 to 18 months. For example: cut release failures in half, build a product team that ships without the CEO in every review, or hire eight senior engineers in a year without lowering the bar. Then list the three or four capabilities that make those outcomes likely.

The team in the Business Insider example did a version of this: before interviews, each interviewer picked a few skills and designed questions to test them, and the final decision was made on scorecards. A scorecard agreed in advance stops a well-rehearsed candidate from winning on style.

2. Structured interviews

Ask every candidate the same core questions, tied to the scorecard, and rate the answers against a shared standard. It can feel stiff, but the evidence backs it. A 2022 re-analysis of selection research in the Journal of Applied Psychology found structured interviews to be the top-ranked selection procedure for predicting job performance (Sackett et al., 2022).

For senior roles I push past the first answer. Ask for a specific decision, then the numbers, then who disagreed, then what they would change. Generated answers are fluent and general. Real experience has names, dates, trade-offs and the odd embarrassing detail.

3. Work samples drawn from your real problems

No puzzles, no personality games, no unpaid weekend projects. Give the finalists a real problem, trimmed of anything confidential, and talk it through live. A QA leader can walk you through how they would build a test strategy for your release process. A talent leader can show you how they would fill your next ten roles. A product leader can critique a real roadmap.

The live conversation matters more than the document. Anyone can produce a clean memo now. Fewer people can defend it when you ask why they cut the thing your CEO cares about.

My view: if the exercise takes more than a couple of hours, pay for it.

4. References and back-channel checks

Done properly, references are the hardest signal for AI to fake. Skip the "would you rehire them?" formality. Ask about the outcomes on the candidate's CV: what was the starting point, what exactly did this person do, what did it look like a year later. Ask what they would need around them to succeed.

Talk to people the candidate did not pick: a former peer, someone who reported to them, someone who worked across from them in sales or finance. My rule is to tell the candidate you do back-channel checks and never to contact a current employer without their consent. If you hire in Europe, check with counsel how data protection law applies to references.

5. Verify identity and history, especially for remote hires

This applies to every remote candidate, in any country. Keep it simple:

  • Cameras on for every interview, with more than one interviewer across the process.
  • Check a government ID against the person on camera before the final stage.
  • Verify employment dates and titles directly with past employers, not only through the candidate.
  • Meet finalists in person at least once for a senior role. Gartner's own guidance on candidate fraud includes in-person interviews as a safeguard.
  • Ship equipment only to a verified address, and make sure the person who shows up on day one is the person you interviewed.

None of this needs to feel hostile. My view is that good senior candidates expect it.

A note on AI screening tools

If you use software to filter or rank candidates, know the rules first. The EU AI Act lists AI systems used "to analyse and filter job applications, and to evaluate candidates" as high-risk (EU AI Act, Annex III). A 2026 amendment moved the main obligations for these systems to 2 December 2027 (Freshfields). New York City's Local Law 144 bars employers from using an automated employment decision tool unless it has had a bias audit within the past year and candidates get notice (NYC DCWP). For a senior search with a few dozen serious candidates, my view is that you do not need a screening tool at all.

Checklist

  1. Write an outcomes scorecard for the first 12-18 months before anyone is sourced.
  2. Build your long list from direct research and referrals. Treat inbound as a bonus.
  3. Use the same core questions for every candidate and score against the scorecard.
  4. Replace take-home puzzles with a real problem discussed live. Pay if it takes more than a couple of hours.
  5. Run at least two references the candidate did not choose, and ask about specific outcomes.
  6. Check ID on camera, verify employment directly and meet finalists in person.
  7. If any software screens candidates, confirm what EU and New York rules require before you use it.

If you are about to hire a senior technology, product or talent leader and want a second opinion on the process, or want me to run the search, you can register the role or send a short brief at www.exectechlabs.lat. I will tell you honestly how I would approach it, including the parts you can do well without me.

This article is general information, not legal advice. Check local rules on background checks, references and AI tools with counsel in each country where you hire.