TL;DR

The resume has become theater. AI writes it, candidates polish it, and a growing share of what it claims is exaggerated or invented. If you hire off resumes and interview vibes, you are scoring a performance of competence, not competence. Replace it with work samples, structured evidence, and a behavioral read of what your actual top performers do, so you can cut mis-hires in half.

Key insights

• The resume signal is degrading. A Resume Builder survey reported by Computerworld in April 2025 found 45% of respondents said they exaggerated skills with AI tools during hiring.
• It will get worse. Gartner projects that by 2028, a quarter of job candidates could be fake.
• The fix is evidence, not better screening of paper. Make candidates do a slice of the real job.
• Benchmark against your own winners. The best predictor of a good hire is what your actual top performers do, not a generic job description.

Why resume theater fails a founder

At 30 people, every hire bends the trajectory. A mis-hire costs months of runway, a manager's attention and a team's morale. Founders know this and still lean on the resume because it is the only artifact they get before the first call.

That artifact is now cheap to produce. Generative AI drafts a tailored resume in seconds, and the Computerworld reporting cited above shows candidates are using it to inflate skills, with some employers reporting that a meaningful share of interviews involve some level of fraud, especially for engineering roles. The resume was never a strong predictor. It is now a weaker one.

What to replace it with

1. A paid work sample

Give finalists a scoped, real problem: a messy dataset, a customer email thread, a roadmap tradeoff. Pay for the time if it is more than an hour or two. You see how they think, how they communicate and whether they ask good questions. A work sample is the one signal an AI-polished resume cannot fake.

2. A structured interview with a scorecard

Decide the three to five behaviors that matter for the role before you meet anyone. Ask every candidate the same questions. Score each answer against the scorecard immediately. This strips out charm and pattern-matching and makes candidates comparable.

3. Reference checks that ask about behavior

Skip "would you hire them again." Ask: when deadlines slipped, what did they do? How did they handle disagreement with a peer? Behavioral questions get behavioral answers.

4. A look at how your winners actually work

This is where most hiring processes stop short. They describe the ideal hire from the job spec, not from the people who actually carry your team. Hatchproof reads digital exhaust from Slack, email, 1:1s and execution patterns, and our behavioral model scores Habits, Aspirations, Temperament, Conviction and Hard skills. Use that to see what your top performers have in common, then hire against that profile. Hatch, our AI agent, can answer "what do my best three engineers do that my median engineer does not?" directly. The goal is half the mis-hires, not a prettier shortlist.

A simple process for a lean team

1. Pick one open role. Write the five behaviors that predict success in it, using evidence from your best people.
2. Screen on a short work sample before any live interview.
3. Run one structured interview with a scorecard.
4. Do two behavioral reference calls.
5. Review the first 90 days against the profile and feed what you learn back in.

Then measure the outcome. We cover how in Quality of Hire: The Only Hiring KPI That Matters, and the broader framework in What is Behavioral Performance Intelligence? See pricing for what fits your headcount and the Trust Center for how we handle consent and data access.

FAQ

Are resumes still useful in hiring?

As a basic filter for work history and role fit, yes. As evidence of capability, much less. AI makes it easy to tailor and inflate claims, and a Resume Builder survey reported by Computerworld in April 2025 found 45% of respondents said they exaggerated skills with AI tools. Use the resume to decide who gets a work sample, not who gets an offer.

What is a work sample test?

A work sample is a scoped task that mirrors real work in the role, such as analyzing a dataset, drafting a customer response or making a roadmap tradeoff. It shows how a candidate thinks and communicates. Keep it to a few hours at most, pay for anything longer than an hour or two, and score every candidate against the same criteria.

How do you spot fake candidates?

Watch for mismatches between the resume and live detail: vague answers about specific projects, reluctance to turn on video, or inconsistent references. Gartner projects that by 2028 a quarter of candidates could be fake, so verify identity and run a live, structured work session. Process beats instinct here.

How can behavioral data improve hiring?

It tells you what success looks like at your company. By reading how your top performers actually work, you can define the behaviors that predict a good hire and hire against them instead of a generic job description. It supports human judgment. It does not replace the interview or the work sample.

Is it fair to use workplace data in hiring decisions?

Use it to define the profile for a role, with consent and transparency, and keep humans responsible for every hiring decision. Audit for bias, avoid penalizing communication styles or time zones, and take advice from employment counsel. This is general information, not legal advice.

Bottom line

Stop grading the resume and start watching the work. Use paper to decide who gets a task, use the task to decide who gets an interview, and use your own top performers as the benchmark. That is how a lean team gets half the mis-hires.

Book a Hatchproof demo