AI candidate scoring
Every profile and every application, checked against your brief
Write the criteria once. Leonar checks each sourced profile, each contact in your database and each CV that applies, then ranks the list. You see every criterion, and you decide who moves on.
- A verdict per criterion
- Applications scored out of 100
- Your talent pool too
- You decide
Used by 400+ high-growth recruiting teams and agencies globally
Sound familiar?
180 applications, 60 sourced profiles, one mandate
The ad went out on several job boards and the applications keep coming. You also sourced sixty profiles for the same role. Somewhere in those two piles are the fifteen people worth calling this week. Candidate scoring in Leonar reads both piles against the same brief, so you start with the right ones.
How it works
Candidate scoring in three steps
One brief per project. The same criteria everywhere.
-
Write the brief
Sort the criteria into must-haves, important, scope, exclusions and nice-to-haves. Leo can draft them from the job description.
-
Leonar reads each profile or CV
Sourced profiles are checked when you confirm a batch. A CV that applies is parsed, then compared with the project brief.
-
Sort by score, open a profile
The list sorts by score. Open any profile to read each criterion, then move the right people forward.
What you get
One way of scoring, in three places
Sourced profiles
Matches, Does not match, Insufficient information
When you confirm a batch of sourced profiles, each one gets a verdict on every criterion. A missing detail is shown apart from a real mismatch, so a strong candidate with a thin profile stays in the list. A detected exclusion caps the score at 30% and names the criterion.
- Batches of 20, 50, 100 or 200 profiles, started by you
- Cancel at any time; finished scores are kept
- The score reflects the criteria that match, not a confidence level
Inbound applications
Every CV scored out of 100 against the brief
With the AI application review switched on, Leonar parses each CV that applies and scores it against the project brief. Each application shows a score out of 100, a recommendation and the verdict on each criterion. Then filter the applications by status, source, project, owner, CV content or experience.
- Recommendations: Strong fit, Fit, Uncertain, Low fit
- Verdicts include Likely (to confirm), for what a call should check
- 0.5 credit per CV analysed; rescoring an analysed CV is free
Your talent pool
The people you already know, ranked for the new role
From a project, Search my talent pool builds the filters from the brief and ranks your own contacts with the same AI score. Pick 20, 50 or 100 to evaluate. With the AI-qualified talent pool module, filter on job family, seniority, availability, notice period, contract, work mode and salary.
- Filters generated from the brief, editable in the sidebar
- Click a score to see what is met, probable or not verifiable
- Contacts already in the project are left out
Before you trust a score
What agencies ask about AI scoring
-
Can a score hide a good candidate?
Less than a keyword filter. Missing information is shown apart from a mismatch, and you can open every criterion before you decide.
-
Does the AI reject candidates for us?
No. Leonar ranks and shows the verdicts. You move people forward or reject them yourself.
-
Can we weight what matters most?
Yes. Must-haves, important criteria and nice-to-haves are separate groups in the brief, and exclusions cap the score.
Pricing
AI scoring is part of the Professional plan
From $149 per user per month, billed annually upfront
- Annual prepaid$149
- Annual commitment, billed monthly$169
- Monthly$199
Per user per month, excluding tax
Professional plan
- AI sourcing and matching, with Leo
- ATS and CRM in one database
- Career page and applications inbox
- Email, SMS, WhatsApp, call and task steps in sequences
- Email and phone finder credits
Good to know
- AI sourcing is activated at setup
- AI application review: 0.5 credit per CV analysed
An all-in-1 tool, essential for nowadays’ recruiters. It supports us in finding the perfect candidates and even became a selling argument to our clients.
Adrien TedjirianCEO, Onos Talents
See your next shortlist scored
Bring a live mandate to a demo: we score real profiles and applications against your brief. Or start your free trial.
The candidate scoring guide
How to read an AI candidate score, how to sort applications from a multiposted job, and how to write criteria a machine can check.
Contents
What is AI candidate scoring?
AI candidate scoring compares a profile or a CV with the criteria of a role and puts a number on the fit. Recruiters use it to decide who to read first, not who to hire.
Scoring profiles is only useful if you can see what is behind the number. A score that says 74% tells you little. A list that says “qualified accountant: matches, Leeds area: does not match, SAP: no information” tells you what to ask on the phone.
Leonar uses one brief per project and the same logic everywhere: for sourced profiles, for your own contacts and for inbound applications.
How to read an AI candidate score
Start with the exclusions. A profile that hits one is capped at 30%, whatever else it matches. The profile shows which criterion caused it, so you can check whether the AI read it right.
Then read the must-haves. “Does not match” on a must-have is a real gap. “Insufficient information” is not: the profile simply does not say. Treat it as a question for your first call.
For applications, the recommendation (Strong fit, Fit, Uncertain, Low fit) gives the overall read, and “Likely (to confirm)” marks what is probable but needs checking.
Where Leonar scores candidates
The same brief drives three kinds of scoring. Each one starts when you ask for it.
| What is scored | How it starts | What you see | |
|---|---|---|---|
| Sourced profiles | Profiles from your CRM and Leonar Source | You confirm a batch of 20 to 200 | A verdict per criterion and a score |
| Your talent pool | Contacts already in your CRM | Search my talent pool, then Evaluate with AI | An AI score column; click it for the details |
| Applications | CVs that apply to a job | AI application review, switched on by an admin | A score out of 100, a recommendation, verdicts |
Sourced profiles
- What is scored
- Profiles from your CRM and Leonar Source
- How it starts
- You confirm a batch of 20 to 200
- What you see
- A verdict per criterion and a score
Your talent pool
- What is scored
- Contacts already in your CRM
- How it starts
- Search my talent pool, then Evaluate with AI
- What you see
- An AI score column; click it for the details
Applications
- What is scored
- CVs that apply to a job
- How it starts
- AI application review, switched on by an admin
- What you see
- A score out of 100, a recommendation, verdicts
Sorting applications from a multiposted job
Switch on the ATS module and the AI review
An admin turns them on in the workspace settings.
Write the project brief
Applications are scored against it. Without a brief, there is no score.
Publish the job
On your career page, or on several boards with job multiposting, activated and billed at setup.
Let the CVs come in
Each CV is parsed and scored out of 100, with a recommendation.
Filter, then call
Filter by status, source, owner, CV content or experience, and start with the strong fits.
Writing criteria a machine can check
Three to five must-haves
Every extra must-have narrows the list. Move the rest to important or nice-to-have.
Write what a CV shows
“Has run a month-end close” can be read on a CV. “Calm under pressure” cannot.
Use exclusions
Current clients, a sector you do not touch. They cap the score at 30%.
Be exact on degrees
Name the qualification you need. A related degree does not prove it.
Keep the grey verdicts
Insufficient information is a call to make, not a reason to reject.
Rescore after edits
Change the brief, then score again. Rescoring an analysed CV costs nothing.
What does the AI candidate score mean?
It reflects how many of your criteria a profile matches. It is not a confidence percentage. Each criterion shows its own verdict, and a detected exclusion caps the score at 30%.
Can I weight the criteria that matter most?
Yes. The brief separates must-haves, important criteria, scope, exclusions and nice-to-haves, so a missing nice-to-have weighs less than a missing must-have.
Does Leonar score applications as well as sourced profiles?
Yes. With the ATS module and the AI application review switched on, each CV that applies is parsed and scored out of 100 against the project brief, with a recommendation.
What happens when information is missing?
The criterion shows Insufficient information, apart from Does not match. Missing information alone never counts as a mismatch.
Does scoring cost credits?
Analysing a CV in the AI application review costs 0.5 credit; scoring it against the brief and rescoring it are free. Evaluating sourced profiles uses 1 credit per 200 profiles.
Can I filter candidates as well as score them?
Yes. Candidate filtering works on titles, companies, location and radius, skills, schools, experience and languages for sourcing, and on status, source, project, owner and CV content for applications.