AI for HR and recruitment, without letting it decide who gets hired

We will not build that, and it is worth being direct about why. Automated screening learns from your past hiring, which means it reproduces whatever bias that contained — and it does so at scale, consistently, in a way that is very hard to defend at a tribunal. Employment decisions are also classified high-risk under the EU AI Act. We automate the coordination around hiring; people decide who is hired.

Summarising for a human reader is more defensible than scoring, but it still shapes what the reader sees, and what it omits is invisible. If you want it, we would build it so the original is always one click away and the summary is never the only thing seen.

For policy and process, yes — retrieval from your approved handbook, with the source shown so the answer can be checked. Anything touching grievance, disciplinary, capability or a protected characteristic routes to a person immediately. Those are conversations, not lookups.

Retention, access and deletion are designed in — candidate data has a defined lifespan and most organisations keep it far longer than they should. We build to your requirements; we do not advise on whether they are correct.

Usually not. Most of this sits around the system you already run, and the common gap is that HR and payroll do not talk, which is an integration problem rather than a product one.

That is your decision, not a technical outcome, and it is worth settling internally before starting. Most organisations we work with are trying to absorb growth without adding administrators, or to move HR time from processing to the casework that actually needs judgement.