Smarter AI Matching

Overview collage of the AI matching refinements in this release: the relabelled slider, extracted-skill badges, and the Tidied/Keywords experience fields

v2.6.0 brought the first user-facing pieces of AI matching online. This release tunes that engine based on the first few weeks of real usage — clearer controls, extracted skills that are easier to see, and cleaner input for the engine to work with.

New to AI matching? See it in action in our short videos.

🎚️ Clearer Slider Labels

The AI matching slider on the New Search screen, labelled Specific skills at one end and Broader experience at the other, with its tooltip open

The slider that balances keyword matching against semantic similarity had labels that described the underlying technique rather than what it actually does for a search. Exact requirements and Related experience are now Specific skills and Broader experience.

The tooltip has been rewritten to match: moving the slider towards Specific skills gives more weight to matching the exact skills and keywords in your requirements text, while moving it towards Broader experience gives more weight to candidates whose overall experience is a similar fit, even if they don’t use the same words. The default stays at the midpoint, weighting the two equally.

🏷️ Extracted Skills, Front and Centre

The AI match requirements field showing extracted skills as a row of badges underneath it

Skills extracted from your AI match requirements text now appear as a row of badges, rather than a plain line of comma-separated text, making it easier to see at a glance exactly which skills the engine has picked out. Those same extracted skills are now also highlighted alongside your typed keywords wherever a candidate’s experience is displayed, so it’s clear which parts of a candidate’s background matched.

📄 Better Job Descriptions for Matching

When a search is linked to a job that has a summary, Talent Catalog now uses that summary as the primary source for the AI match requirements, instead of combining it with less structured intake and job-description-file text. If the job has no summary, its requirements are still built from those other fields as before. Candidates with no genuine match are also no longer returned — previously they could appear at the bottom of the results with a zero match score.

You’ll also see a loading indicator while job-matching information is being fetched for a search, so it’s clear that something is happening rather than that the screen has stalled.

🧹 Cleaner Experience Data

Admin portal job-experience edit form showing the Original, Tidied, and Keywords description fields side by side

Candidate job experience descriptions can now have two extra fields alongside the Original text: a Tidied version, and a comma-separated list of Keywords. These are visible and editable from the admin portal. Formatting such as HTML is also now stripped from text before it’s used for skills extraction, so it no longer adds noise to matching. The candidate portal’s own experience form is unchanged — candidates still see and edit a single description field, just as before.

📤 Export and AI Matching

CSV export is no longer offered for AI-matching searches. Export didn’t properly support matching searches, so rather than risk a download that didn’t reflect what was on screen, the option has been turned off for these searches until that work is revisited. Exporting an ordinary, non-AI-matching search is unaffected.

🚀 What’s Next