How should our leadership team redesign our candidate evaluation process when AI allows applicants to fake technical assessments and portfolio work?

Question

Tyler Smith · Accepted Answer

If your hiring process relies on take-home writing prompts, static code tests, or polished portfolio presentations, you are no longer evaluating the applicant. You are evaluating how effectively they prompt generative software. You must modernize your hiring screen immediately.

In EOS, we evaluate people against the Accountability Chart using GWC: do they Get it, do they Want it, and do they have the Capacity to do it? When AI handles the mechanics of job execution, measuring Capacity requires testing real-time problem-solving and critical judgment, not rehearsed deliverables.

Shift your hiring plan in three practical ways:

- Eliminate take-home tests: Replace unsupervised assignments with live, collaborative working sessions during the interview. Present a messy, real-world business scenario and ask the candidate to talk through their diagnostic reasoning.
- Probe the edges of AI output: Give the candidate an AI-generated deliverable loaded with subtle factual inaccuracies, logical gaps, and compliance flaws. Ask them to audit the deliverable, spot the errors, and explain how to fix it. This tests true subject-matter depth.
- Test prompt agility: Have the candidate work alongside an AI tool live in the room. Observe whether they blindly accept the first output or steer the tool with precise, iterative instructions to produce a superior result.

Hire for critical thinking, emotional intelligence, and domain mastery. Tools will change every six months, but people who truly Get It and have the Capacity to direct those tools will drive enterprise value.