AI and automation for recruitment agencies
We build AI for recruitment agencies, staffing firms and in-house talent teams, with recruitment automation throughout: CV parsing into your ATS, shortlists that explain themselves and back-office flows from placement to payroll. Recruiters make every selection decision, and the system records why a candidate was put forward.
Where the work gets stuck.
Admin before the first conversation
Recruiters spend hours reading CVs, updating Bullhorn or Carerix and chasing references before they speak to a candidate. The database grows while its records go stale.
A database that is hard to search
Keyword search misses the candidate who did the job under a different title three years ago. Agencies pay for job board access to find people already in their own ATS.
Placement to payroll by hand
Once a flex worker starts, contracts, hours and invoices move between the ATS, the payroll system and finance, often through spreadsheets. Mistakes surface on a payslip or a client invoice.
How we improve recruitment.
Each one is scoped around your systems and rules, and each one keeps a person in charge of the decisions that matter.
CV parsing and profile enrichment
Incoming CVs and application emails become complete candidate records with skills, experience and availability, checked against duplicates already in the ATS. Recruiters correct the record in one screen before it is saved.
Matching across your own database
Recruiters describe a vacancy in plain language and get candidates from the ATS with a short explanation of each match, including people with different job titles. The recruiter decides who to contact, and no candidate is rejected automatically.
Vacancy intake and job ads
Notes from a client intake call become a structured vacancy and a first job ad in your house style, checked for discriminatory wording. The consultant edits the ad and publishes it.
Candidate scheduling and follow-up
An assistant handles availability checks, interview scheduling and reminders by email or WhatsApp, and hands the conversation to a recruiter when a candidate asks something substantive. Candidates are told they are talking to an AI assistant.
Placement to payroll
A signed placement creates the contract, hours registration and invoicing setup in Mysolution or Easyflex, with rates checked against the client agreement and the applicable CAO scale. Payroll staff approve anything that falls outside the rules.
Client lead qualification
Enquiries from companies looking to hire are enriched, scored against your ideal client profile and booked with the right account manager. Qualified leads hear back quickly, and the rest get a useful reply.
Built around the rules.
What we design for from the first week. Your legal and compliance people keep the final word.
EU AI Act high-risk requirements
AI used to screen, filter, rank or evaluate candidates is a high-risk use case under the EU AI Act, with duties on human oversight, logging, documentation and informing candidates. We design matching so recruiters decide and every ranking is logged with its reasons. Emotion recognition in recruitment is prohibited, and we do not build it.
GDPR for candidate data
Candidates can access, correct and delete their data, and the Autoriteit Persoonsgegevens expects data of rejected applicants to be removed within a few weeks unless they consent to longer retention. We build retention and consent rules into the database and avoid solely automated decisions under Article 22.
Waadi registration and the Wtta
Staffing firms must be registered under the Waadi, and the Wtta admission system for labour providers starts on 1 January 2027. Placement and payroll automation keeps clean records of pay, hours and hirer agreements under the ABU or NBBU CAO you apply, which audits and admission checks ask for.
Equal treatment in job ads and matching
Dutch equal treatment law applies to how vacancies are written and candidates are selected. We test job ad drafts for biased wording and check matching results for uneven outcomes across groups before release.
Works with what you run.
If a system has an API, a database, an export or an inbox, we can build on it. These are the ones we meet most.
- Bullhorn
- Carerix
- Mysolution
- Recruitee
- Homerun
- Textkernel
- LinkedIn Recruiter
- Easyflex
- Salesforce
- Microsoft Outlook
Where to start.
CV parsing into your ATS
We start with the intake of new applications for one office or specialism: CVs and emails parsed into complete ATS records, duplicates flagged and profiles summarised for the recruiter. It removes typing work early and gives later matching work a clean database.
Talk it throughWhat it includes
- Mailbox and job board intake
- Parsing into Bullhorn, Carerix or Mysolution fields
- Duplicate detection
- Recruiter review screen with logging
Related work.
Guides.
AI readiness checklist: what to answer before your first agent goes live
The questions to answer before your first AI agent handles real work, from process and data to approvals, monitoring and ownership.
10 min read
AI vendor security questionnaire: what to ask before you buy or build
The questions to ask an AI vendor, or your own team, about data, model providers, access, logging, quality, incidents, GDPR and exit, and how to score the answers.
8 min read
Further reading.
Human-in-the-loop AI agents: design patterns for production
The patterns we use to run AI agents safely in production: approval checkpoints, confidence thresholds, tool permissions, audit trails and fallbacks.
4 min read
Embeddings and semantic search: a practical guide for product teams
How embeddings power semantic search, how to choose a model and a vector store, and how to measure whether your results actually improved.
5 min read
LLM evals: how to test AI features before every release
A practical approach to LLM evals: build a test set from real cases, combine code checks with model grading, and block releases that regress.
5 min read
Common questions.
Yes. AI systems that screen, rank or evaluate candidates fall into the high-risk category, which brings obligations on human oversight, logging, transparency to candidates and monitoring. We build those into the design and documentation, and help you work out whether you act as provider or deployer for each tool.
No. Our matching tools rank and explain, and a recruiter decides who is contacted, put forward or turned down. Keeping that decision with a person also keeps you clear of solely automated decisions under the GDPR.
Yes. Both offer APIs for candidates, vacancies and placements, and we build on their data models. Where a staffing system has a limited API, we agree the approach during discovery.
Textkernel is strong at parsing and search inside the systems that support it. We add the steps around it: reading application emails, cleaning duplicates, drafting summaries and linking the result to placement and payroll. If your current setup already covers the need, we say so.
Before release we run matching against historical vacancies and check whether results differ across groups in ways the job requirements do not explain. The same checks run after every change, and recruiters can flag rankings that look wrong.