The starting point
The company had a database of more than 50,000 past candidates in its ATS (Applicant Tracking System). Hiring managers had access to it, but search worked only on exact keywords.
The problem
- Poor search. If a manager used the wrong keyword, they found nobody. If they used one that was too broad, they got thousands of irrelevant results.
- Wasted potential. The company had paid to acquire those candidates in the past but could not search them because the system was bad.
- Frustrated managers. Rather than finding people themselves, managers had to keep asking the recruiting team for help.
The solution: an internal AI sourcing tool
We built a system that completely changes how managers search the internal database. Instead of keywords, they use natural language.
How it works:
- Search in plain language. The manager writes who they are looking for (for example “I need a senior backend developer with payment gateway experience and Python”).
- AI analysis and ranking. The system connects to the ATS through its API, runs several searches in parallel, analyses candidate profiles and ranks them by how well they match the request.
- Dynamic filtering. While reviewing results, the manager can add further criteria on the fly and the system reorders the candidates immediately.
What changed
- Speed. Managers find relevant candidates in a fraction of the time.
- Accuracy. The AI understands context. If a manager asks for “payment gateways”, the system also surfaces a candidate whose CV says “Stripe” or “Braintree”, even though they never wrote the phrase.
- Independence. Hiring managers are far more self-sufficient, and the recruiting team has more time for the interviews themselves.
What transfers to your company
- Use the data you already have. Many companies sit on a goldmine of historical data (candidates, old enquiries, documents) but cannot search it. AI changes that.
- Natural language is the new interface. If your people have to master a complicated search syntax, the system is bad. AI lets you ask systems the same way you would ask a colleague.
- Semantic search. Traditional search matches words. AI matches meaning. That is an enormous difference in the quality of the results.
How to bring this into your company
If you want to know what search like this over your own data would look like, let us start with a 15-minute call. We go through whether it makes sense in your company and which of our workshops is the best starting point.