AI for Recruitment
Agencies
Read the CV and notes
Check the role
Update the CRM
Consultants keep control while the admin gets done.
Agents for a recruitment desk
CV processing: This agent receives CVs in a variety of formats and turns them into a nicely-formatted, branded, Word CV, which is passed on to the relevant consultant to approve. Only then does it get shared with clients.
Spec matching: Consultants upload new roles to your CRM. This agent matches the role to relevant candidates already in your CRM, returning a prioritised, searchable, and shareable shortlist with reasons for inclusion and the identities of any consultants who’ve placed them previously.
Candidate screening: This agent follows your screening process. It can be built to document decision-making and provide an audit trail for ICO or insurer requests. It can include candidate and consultant-facing processes for appeal, fairness, and data privacy.
Candidate checks: This agent conducts candidate checks before video interviews, including identity, sanctions, right-to-work, IR35, etc. It flags inconsistencies and potential fraudulent CVs.
First-line candidate interactions: Using voice and/or chat, this agent screens candidates and confirms interest and availability, sends calendar invites, requests references, etc., then hands off to consultants with transcripts.
How would the solution be implemented?
Book a working session with us to map your team’s workflows, identify repetitive tasks suitable for automation, and discuss any compliance or efficiency pain points. Ideally, we’d shadow a consultant for a day. From there, we’d rank workflows by billing impact, ease of implementation, and compliance risk, then discuss next steps.
We build on top of your recruitment CRM, such as Bullhorn, Vincere, etc. Wherever possible, we’ll make use of native capabilities within your platform. Custom solutions will be built where your use-case demands, including your branding and CV templates, industry-specific fields, historical candidates, regulatory requirements, etc. We’ve done this already in our work with AM2PM, so it’s an area where we have experience.
The solution relies on a company-wide data layer to provide agents with access to relevant information, such as candidate data stored in the CRM, job specs, CVs, etc. This layer would include controls and governance, including role-based access, source citation, human-approved data only, etc.
The solution would start small, focusing on a particular use-case, such as candidate screening, CV processing, or dormant database activation, and/or limited scope such as a handful of perm/contract/executive search/RPO desks, etc. Your team would test outputs, including CRM owners and relevant desk consultants, and we’d discuss wider implementation.
As new workflows are added, we’ll document everything, build in human oversight, and case-test to provide an audit trail should ICOs or insurers come knocking.
How much will it cost?
The audit is a fixed fee. Following the audit, you’ll receive a quote for each agent, phase-scoped if necessary, including success criteria and implementation timescales. We’ll agree these figures before we write a line of code. Following implementation, you can engage us on a retainer basis or not, which is your choice.
What about?
EU AI Act compliance? Some recruitment systems sit on its high-risk list, and those rules don't apply yet, but your UK data protection and equality duties apply now.
AI bias? There’s no such thing as zero bias, but the solution can be built with mitigating factors, such as criteria tied to the job spec, disparate impact tests for legitimate inequalities, human oversight, candidate-facing appeal process, etc.
UK equality and data privacy regulations? We adhere to these today, so they’ll be incorporated into any AI-enabled process we create for your agency.
Buy-in? Consultants hate admin. Provided review/approval requirements don’t outweigh time savings, they’ll likely welcome agents to take on drudge work. We’ll consult with relevant desks on use-case selection to ensure ownership.
Fudged CVs? We don’t check if they’re AI-enabled (yet) but process them as normal, extracting relevant information around experience, skills, seniority, timeline, etc. If consultants flag a concern, they’ll be rejected just like any other fraudulent CV, ideally before Day 1, not after.
How is that different from life as we know it?
Right now, consultants stay late to format CVs. Your candidate database is a rounding error in terms of utilisation. Candidate screening and compliance is left to the fickle memory of individual consultants. CV fraud isn’t flagged until a candidate’s Day 1.
Get in touch
Book a working session to learn more about how we can bring your agency’s CRM to life.