Matching

Matching reverses the usual job board. Instead of a person hunting for jobs, you hold the talent and talentkiwi hunts the market: one candidate goes in, and a ranked shortlist of fitting jobs comes out, each with a 0 to 100 fit score and a plain-language explanation of why it fits.

Before you start

Matching always begins with a candidate. Create a candidate project and upload the CV first, plus any recruiter notes or extra files. talentkiwi reads the documents automatically, including scanned PDFs via OCR, and pulls everything into a single profile the AI can work with.

Preview the strategy before it runs

Before committing, you see a Search Strategy Preview: the AI explains in plain language how it interpreted your request and which filters it's applying. A live counter shows how many jobs currently match (for example, "searching in 1,240 jobs"), so you can widen or narrow the search before spending the run. Refine and re-preview as many times as you like.

Define what you're looking for

Shape the search with three controls:

• Sources — pick where to search: the broad public job market, your own company data or ATS, and more.
• Query — describe the profile or role in your own words. You can type it or dictate it by voice.
• Smart filters — exclusion rules such as experience level, workload (%), languages, location, and company. The AI proposes a sensible starting set; adjust it freely.

For location, you can search by place plus radius, by region or canton, or by real travel time (for example, within 30 minutes by car or public transport). Commute time is then calculated for every job.

How the matching works

When you start the search, the engine works in layers to find true fit rather than keyword overlap:

1. It reads for meaning, so it surfaces relevant roles even when the wording differs from the CV.
2. It combines that with precise keyword matching for accuracy.
3. It re-ranks the best results by genuine relevance.
4. It scores each job from 0 to 100 and writes a short explanation of the fit.

A progress bar keeps you informed while it runs.

Review the results

Each job comes back as a card showing:

• A color-coded fit score ring (green for excellent, amber for fair, and so on).
• Job title, company, location, and a New badge for fresh matches.
• Recruiting detail: commute time, workload (%), languages, posting date, distance, and a flag for agency-posted listings.
• The AI fit explanation in two panels, "Why it fits well" and "Areas to improve", so you get the reasoning, not just a number.
• The full job description and a link to the original posting.

You can re-rank results by best match, newest, distance, or match date, and apply a minimum quality threshold or filter by commute time. When a posting states that staffing agencies are not wanted, you can sort those jobs out. All adjustments apply instantly, without a new search.

Act on matches

For each job you can:

• Shortlist or reject it. Results are organized into Suggested and Shortlisted tabs.
• Create a document project pairing this candidate with this job in one click. talentkiwi tells you if one already exists, so you never duplicate work.
• See the direct contact for the role (name, phone, email), all copyable.
• Open the original posting.

You can also work in bulk: shortlist, reject, or create projects for many jobs at once, and generate email drafts to present jobs to your candidate or present your candidate to a client.

Alerts and refreshing

Saved-search alerts — turn any search into a standing alert that runs on its own and emails you new matching jobs as they appear, as a morning digest, hourly, every few hours, or weekly. Stale alerts pause themselves so your inbox stays clean.

Additive refresh — after a few days, re-run a search to pull in only the genuinely new postings. Your reviewed and shortlisted jobs stay untouched, and new finds are flagged.