AI in recruitment, how SMEs keep hiring fair
AI can save SMEs time in hiring, but it can also quietly make things less fair if it becomes a black box. This guide shows where AI helps, where it harms, and the simple guardrails that keep you accountable. It also includes a practical process you can copy, even without an HR team.
Written by: Frankie Brookton, Founder, inyourroots®
AI can save SMEs time in hiring. But it can also quietly make the process less fair if it becomes a black box.
You do not need to ban AI. You just need to use it in a way that keeps hiring fair, explainable, and accountable.
Where AI helps (and where it quietly harms)
AI can help with:
- drafting job ads in plain language
- organising applications
- summarising notes and interview feedback
- scheduling and admin
AI can quietly harm when it:
- screens people out based on keywords they do not know to use
- rewards “polish” over potential
- makes decisions that no one can explain
- creates a false sense of certainty
Simple rule: use AI to reduce admin, not to replace judgement.
The “black box” problem (and how to avoid it)
A black box is when:
- the candidate does not know what is being assessed
- you cannot explain why someone was rejected
- nobody is clearly responsible for the decision
To avoid it, keep three things true:
- Transparency: you can describe the criteria in plain English
- Human ownership: a named person makes the final call
- Appeal path: you can review edge cases, especially for first-timers
Simple guardrails (you can implement this week)
- Do not upload sensitive personal data into tools you do not trust
- Do not use AI as the final decision-maker
- Use structured criteria, not “vibes”
- Check accessibility, especially for neurodivergent candidates
- Keep a feedback step, even if it is short
Strengths-first screening (without excluding first-timers)
If you want to widen access, design your first step around strengths and proof.
Instead of “send a CV and cover letter”, try:
- a short form: “What are you good at, and what do you want to learn?”
- a small task that mirrors the real work
- a portfolio option (optional): photos, examples, short notes, a short video intro
If you use AI to help you review, use it to summarise, not to score.
A fair process you can copy
- 1.Write the role for strengths (what good looks like, what can be taught)
- 2.Offer a fair first step (task or short form)
- 3.Use a structured interview (same questions for everyone)
- 4.Make a human decision (with notes you can explain)
- 5.Give short feedback (one strength, one improvement)