The AI Illusion
It was 7:52 pm and Our AI had just “interviewed” two candidates.
Candidate A was on fire: perfect camera, perfect background, fluent with terms like “distributed systems,” “circuit breakers,” “backpressure,” “tracing,” “SLOs,” “error budgets,” “blue-green,” “canary,” “Kubernetes,” “Terraform,” “Kafka,”and “DLQs.” Textbook.
Candidate B fumbled for a moment, then started asking real questions when the interview hit a STOP.
Viewed through an AI lens trained on tone and keywords, Candidate A wins and Candidate B loses.
We reached out to Candidate B later and were amazed by the depth of their expertise.
Many tools in the market offer AI video interviews that analyze tone, expressions, and responses to assess communication skills, confidence, and suitability for a role. That signal can be useful, but it is still surface.
What is really happening in the recruitment industry right now:
- Resumes and answers are increasingly AI-polished, so it covers the surface AKA Pre-Screen.
- Feeding Skills-first keywords in AI can definitely expands your reach by 10x for sure. But, when the final interview falls flat, the cycle resets and you chase the “perfect candidate” again.
- Nearly half of applicants now use AI to craft resumes or assessments, which raises volume and SAMENESS.
- Some surveys report over 50% of companies use AI in their hiring process, with projections reaching 68-70% by the end of 2025.
Use AI where it helps
AI is reshaping how we source, sort, and schedule. It can scan large pools, tidy notes, and speed up busywork. AI is a capable assistant, but overuse can dull your signal. Use AI when you need to screen a lot of candidates and your calendar is packed. However, Keep an eye on time-to-hire. If it keeps rising, you are likely optimizing for volume, not outcomes.
So how do we solve this?
- Add Human Curation: AI can de-duplicate and cluster similar profiles. Then a recruiter should reach out and run a short discovery call before panels. This prevents last-minute surprises and wasted interviews.
- Cap the searchLet AI lift the heavy search, but stop at 30 profiles. If you have more, understand that the brief is fuzzy.
- Define your own “must-haves”Do not rely on generic checklists. Write 3–5 must-haves tied to your 90-day outcomes. Everything else is just nice-to-have.
- Interview for culture and collaborationNot just tech. Add a short, role-relevant simple culture pass on pace, feedback style, documentation habits, and cross-team work.Use a 0–2 rubric for each so notes stay objective.
- Close the loopShare clear decision notes, run two references for outcomes and trade-offs (not adjectives), and keep the join warm with a manager’s hello within 72 hours of offer.
Closing thought
AI is great at setup. Humans are great at sense-making. Blend both with care and you will interview fewer people, decide faster, and make hires that stick. Keep the process intuitive and kind. Let tools lift the load. Let people make the call.