AI for recruitment agencies, done fairly.
AI can take hours out of sourcing, screening and admin. It can also reject good candidates, bake in bias and breach the new rules on automated decisions without anyone noticing. We help your agency get the first without the second.
Where AI saves time.
The work that fills a recruiter’s week, and what AI can help with, with a person making the decisions that matter.
- Job adverts
- Clear, inclusive adverts drafted from the client’s brief, checked for language that could put people off applying.
- Searching your database
- Candidates you already know, found by skills and experience rather than keywords, for a consultant to review.
- Screening support
- CVs summarised against the role’s requirements, so consultants spend their time on judgement, not reading.
- Candidate communication
- Updates, reminders and interview scheduling handled promptly, so nobody is left waiting.
- Client submissions
- Candidate summaries and interview notes written up for clients, ready for a consultant to check.
- Compliance admin
- Right-to-work documents, references and timesheets chased and checked.
What to get right first.
Fairness
AI learns from past hiring decisions, including their bias. If it disadvantages people with a protected characteristic, your agency is liable, whoever built the tool.
Automated decisions
If AI ranks, filters or rejects candidates with no real human review, the law treats it as an automated decision, with rules you must follow.
Candidate data
Collect what the role needs, keep it only as long as you should, and know where your tools send it.
The AI already in your recruitment software.
Recruitment CRMs and job platforms are adding AI that ranks, screens and messages candidates. It can save hours a day. It can also make decisions about people without anyone noticing, which is exactly what the law now regulates.
- Bullhorn
- Bullhorn Amplify ranks candidates by relevance, runs AI first-round screening and, since 2026, offers “digital workers” for screening, outreach and prospecting.
- Other recruitment CRMs
- Vincere, JobAdder, Mercury and others are adding AI matching, parsing and messaging. Ask what’s switched on, and whether any of it decides on its own.
- Job boards and LinkedIn
- AI that suggests candidates and drafts messages. Useful, but it brings someone else’s view of who “fits”.
- General assistants
- ChatGPT, Claude and Copilot for job adverts, summaries and client proposals, used through business accounts. See our comparison.
Features vary by plan and change often. Checked November 2026.
Before AI touches a candidate decision
Ten checks for any AI that ranks, screens, scores or rejects candidates, in your CRM or anywhere else.
- Does it decide, or suggest? List every point where AI ranks, scores, filters or rejects candidates.
- Is the human review real? The ICO says a person must genuinely be able to change the outcome. Rubber-stamping counts as an automated decision.
- Do candidates know? Tell people when AI is used, and how to contest a decision and ask for a person to review it.
- Have you done a DPIA? A data protection impact assessment is generally needed before AI screening.
- Could it discriminate? Check outcomes by protected characteristic. You’re liable under the Equality Act, whoever built the tool.
- What does it learn from? Historical hiring data carries historical bias.
- What does it collect? Only what you need for the role, not a profile for ever.
- How long is data kept? Set retention, and check the supplier follows it.
- Are clients’ requirements fair? AI will faithfully apply a discriminatory brief. Challenge it first.
- Is it in your AI policy? Name each approved feature and who oversees it. Our AI policy template is a start.
What the law and the ICO say.
Since February 2026, the Data (Use and Access) Act 2025 has replaced the old near-ban on solely automated decisions with a set of safeguards. You can make significant automated decisions about candidates, as long as you tell them, let them contest the decision, offer a human review and can explain it. The ICO’s March 2026 report on automated decisions in recruitment found many employers didn’t realise they were making them, and was clear that human involvement only counts if it can genuinely change the outcome. CV filtering, suitability scoring and behavioural assessments are the examples it gives. Final ICO guidance is expected this winter.
The Equality Act 2010 applies as it always has: if an AI tool puts people with a protected characteristic at a disadvantage, your agency is responsible, intended or not. Earlier, the ICO’s audit of AI recruitment tool providers made almost 300 recommendations, after finding tools that could filter out candidates by protected characteristic, or collected far more data than needed.
The REC’s guidance on AI GOV.UK: Responsible AI in recruitment
Who you’ll work with.
James has over 25 years' experience delivering technology for large organisations across the public and private sectors, including Critical National Infrastructure programmes for UK Government and major IT programmes for FTSE 100 companies.
Every engagement is led by James personally. He has worked on programmes where security and assurance are non-negotiable, and brings the same rigour to the tools your consultants use.
Start with an AI health check.
A fixed-price review of how your agency uses AI today, and what to change.
- The AI in your CRM and job tools: what’s switched on, and where it ranks, screens or decides
- The tools your consultants use, including the ones nobody signed off
- Automated decisions and fairness: your processes against the new rules and the ICO’s expectations
- Clear next steps, in a short written report that’s yours to keep
£750 + VAT, fixed price.
Questions recruitment agencies ask us.
Can we use AI to screen CVs?
Yes, with safeguards. If AI ranks, filters or rejects candidates without a person who can genuinely change the outcome, it’s an automated decision: candidates must be told, and be able to contest it and ask for human review. You’ll generally need a DPIA and bias monitoring too. Many agencies keep AI to summarising and suggesting, with consultants deciding.
Is our agency liable if an AI tool discriminates?
Yes. Under the Equality Act 2010 the agency is responsible for its recruitment decisions, whoever built the tool. A supplier’s bias testing, often done on different candidates from yours, doesn’t move that responsibility.
Do we need to tell candidates we use AI?
If AI makes significant decisions about them, yes, along with how to contest them. Even when it only assists, saying so in your privacy notice and candidate communications builds trust.
Will it work with our CRM?
Often, yes. Many recruitment CRMs now have AI built in or connect to other tools. We check what yours can do, and what’s already switched on, before recommending anything new.
Do we need an IT department?
No. We work with whoever runs the processes today, and recommend tools your team can look after themselves.
Book a free 30-minute intro call.
Tell us what you're working on. You'll get an honest view of what would help and what a sensible first step would be. No obligation.
Or email hello@contextric.com