Plain-language summary.
We use language models in two narrow places: to classify briefs against our ICP, and to draft a first-pass shortlist for human review. Every shortlist is reviewed and signed off by a human operator before it reaches the buyer. We publish a bias and fairness audit every quarter with the full methodology and results.
1. Where we use AI
AI is used in two operational steps. First, an inbound brief is classified against our Ideal Customer Profile to determine whether it is in-segment. Second, an AI-drafted first-pass shortlist is generated from the vendor pool based on the role specification. Both outputs are advisory only.
2. Models in use
At the date of this notice, we use OpenAI GPT-4-class models for classification and shortlisting, accessed via the official API with zero data retention enabled. No brief data is used to train any third-party model. We disclose the model in use on this page and update this notice within 30 days of any change.
3. Human review
No AI output reaches a buyer without a human operator's sign-off. The operator reviews the AI-drafted shortlist against the brief, makes additions and removals, and accepts or rejects each candidate individually. The final shortlist sent to the buyer is the operator's shortlist, not the model's.
4. Bias and fairness audit
Every calendar quarter we audit the AI-assisted shortlists for bias on protected attributes (where lawfully observable) and for systematic preference for any vendor type, geography, or seniority band. The audit is conducted by an independent reviewer and published in full.
Download the latest audit summary (PDF)
5. Your rights
Under the EU AI Act and GDPR Article 22, you have the right to request a human-only review of any decision that affected your brief. Email ai-notice@xpertdirect.io and a senior operator will re-review your brief without any AI assistance and provide a written response.