AI candidate scoringThe strongest applications first, with the reason why.

Every profile of your project gets a colour and a one-line reason, so you can explain the order to a client.

  • A score on every profile, without a click
  • A written reason you can read out to a client
  • Nobody set aside without you

Used by 1,200+ agencies and recruiters in 25 countries.

The scale

A candidate scoring scale with three levels and a reason

A score out of 100 asks you to guess what separates 84 from 87. Three levels match the three things you actually do with a profile, and the written line carries the detail.

Green, a strong fit

The profile meets the brief on skills, experience and domain. A junior whose degree clearly matches the job can land here too.

You open these first.

6y exp, FP&A in food manufacturingJunior, MSc Finance and Accounting5y exp, SAP consulting

Amber, a fit with gaps

A real candidate with something to check, such as a missing skill, fewer years than asked, a job title that can mean two things or a CV without dates.

You call to check the gap.

3y exp, lacks business partneringFinance exp, missing dates3y Area Mgr, role unclear

Vermillion, a blocker

Kept for two cases only: a language the brief requires that the profile clearly lacks, or a completely different trade. A CV that simply lists no languages stays amber.

You check it before you set them aside.

German required, basic level onlySales career, no finance experience

When there is no score yet, the cell tells you why

Evaluating...

The score is being written, usually while you look at the list.

Incomplete offer

The brief is missing its description, profile sought or location.

Incomplete profile

The profile has no experience or education to read yet.

Incomplete data

The profile was skipped because there was not enough to score.

The three colours come from a palette chosen to stay distinct for colleagues with colour blindness, and the written reason always says the same thing in words.

When the client asks

Explainable AI scoring you can talk through with a client

When a client wants to know why someone is first, or why someone is missing, the reason is already written on the profile. You explain it in your own words and add what you learned on the phone.

Your client, Elena

Why did you call Anna first?

What Jarvi wrote on the profile6y exp, FP&A in food manufacturing
You

Anna has six years of planning work in food manufacturing, which is the closest match to your brief, so she was the first person I spoke to.

Your client, Elena

Is Sara really ready for this role?

What Jarvi wrote on the profile3y exp, lacks business partnering
You

Sara is a little short on business partnering, so I checked it with her on the phone before sending her profile, and she already presents the monthly numbers to two site managers.

Your client, Elena

Marco looks strong. Where is he based?

What Jarvi wrote on the profile8y exp, matchesBristol
You

Marco fits the role well and lives in Bristol. A second remote day each week would make the commute work for him, so I wanted to ask you about it first.

Your client, Elena

Why is Lukas not on the list?

What Jarvi wrote on the profileGerman required, basic level only
You

Your brief asks for fluent German and his CV lists it at a basic level. I have asked him whether that is still true, and if his German has moved on, he goes straight onto the shortlist.

Applicant scoring, step by step

How applicant scoring works on a real project

You keep the way you already work on a role. The score simply tells you where to start and leaves the reading of each profile to you.

  1. Step 1About 2 min

    You write the brief once

    The score reads three fields of the project: the job description, the profile you are looking for and the location. If you already have the job ad, you can ask Jarvi AI to fill them in for you.

  2. Step 20 min, done by AI

    Every profile gets a colour and a reason

    You open the Profiles tab and the AI scoring column fills in on its own, for people who applied and for profiles you added yourself. Each colour comes with one written line that says why.

  3. Step 3

    You open the strongest profiles first

    You start with the green ones, and the amber reason tells you what to ask on the call. When a CV is updated, its score is written again on its own, and when the brief changes you refresh the whole project from the menu next to any score.

Every profile gets its chance

Fair candidate scoring, even when the CV is incomplete

Many good candidates write a short or imperfect CV. The score is built to read the career behind it, and when in doubt it chooses the middle level rather than the lowest one.

Details most CVs never mention

A driving licence, the right to work, office software or availability are rarely written on a CV. When the brief asks for one and the profile is silent, the score leaves it out instead of counting it as a gap.

Juniors with the right degree

A graduate is scored on what they studied, and a degree that clearly fits the job can earn a strong fit.

Living far from the office

Beyond 100 km, for a role that is not mostly remote, the city is added to the reason. The colour stays the same, because the commute is a conversation for you and your client.

A career across several employers

Three years in one firm and two in another add up to five years in the same trade.

A CV without dates

Missing dates make the profile one to check on the phone, never one to drop.

Skills that come with the job

A finance manager who never wrote "budgeting" still did it, so the score does not mark it as missing.

See it on your own roles

Bring a role you are working on and we will score its profiles with you

In a demo, we fill in the brief together, let the profiles get their scores and read the reasons line by line, so you can judge them on a real case.

Honest about the limits

AI-assisted candidate evaluation where you keep the final word

The score is there to save you reading time on a busy role. Everything that follows, from the call to the shortlist, stays your decision.

It never sets anyone aside for you

Nothing is archived, rejected or moved in your pipeline because of a score. The status of each candidate stays the one you choose.

It does not reorder your list

The colours sit in their own AI scoring column, so you scan the project and open the green ones first, in the order you prefer.

It reads the brief, not the dropdowns

The Education Level field is not used. If a degree matters for the client, a sentence in the description such as "a finance degree or equivalent" does the job.

It needs something to read

A brief without its description, profile sought or location shows "Incomplete offer". A profile with no experience and no education shows "Incomplete profile" until you complete it.

Choosing a method

AI candidate scoring compared with filters and match percentages

Every agency already sorts its applications somehow. Here is what each way of doing it gives you on a busy role.

On a busy roleA keyword filterA match percentageReading every CV yourselfJarvi AI scoring
Shows you which profiles to open firstOnly on exact wordsYesAfter the readingYes, by colour
Explains each verdict in wordsNoRarelyIn your headYes, one line
Treats a detail missing from the CV as unknownNoRarelyYesYes
Keeps distance out of the verdictOften filters on itOften counts itYesYes, shown as a flag
Your time on 60 applicationsA few minutesA few minutesSeveral hoursNone, it runs on its own
Leaves the decision to youHides the restUsuallyYesYes

AI candidate scoring FAQ

What is AI candidate scoring in Jarvi?

It is a score that Jarvi gives each profile of a project, measured against the brief of that project. You see one of three colours with a short written reason, so you know which profiles to open first and why. It works on applicants and on profiles you added to the project yourself.

Why three levels instead of a score out of 100?

The difference between 84 and 87 is hard to explain to a client, and you would rarely act differently on either. Three levels match what you actually do with a profile: open it first, call to check a gap, or check the one point that blocks it. The nuance lives in the written reason rather than in a number.

Can the score reject a candidate on its own?

No. The score is information in a column and on the profile. It never archives, rejects or moves a candidate, and the lowest level is kept for two cases only: a language the brief requires that the profile clearly lacks, or a completely different trade. A CV that simply lists no languages stays at the middle level, and whenever there is doubt the score leans towards the middle level rather than the lowest one.

What do I need to fill in for the scoring to work?

The project needs its job description, the profile you are looking for and its location. Each profile needs at least one experience with a job title and a company, or one education. The Education Level field is not read, so if a degree matters, write it in the description.

How are juniors and candidates who live far away scored?

A junior is scored on their education, and a degree that clearly fits the job can earn the top level. A candidate who lives more than 100 km away from a role that is not mostly remote keeps their colour, and the city is simply added at the end of the reason so you can discuss the commute with them and with your client.

In what language are the reasons written?

In the language of the job description. A brief written in English gives English reasons, and a brief written in Dutch gives Dutch ones, which helps when your team works on roles in several countries.

What happens when the brief or a CV changes?

When a profile is updated, its score is written again on its own, with a small spinner on the sparkle while it refreshes. A change to the brief is not picked up by itself, so when the client changes it you choose "Refresh for entire project" in the menu next to a score, and the profiles are scored again against the new version.

Is scoring the same as matching?

They work together. Scoring tells you how well each profile of a project fits its brief. Matching starts from the brief and looks for the right people in your whole database, and the profiles it brings back come with a colour and a reason as well.

See your own candidates scored against your own brief.

In a demo, we score one of your real roles and go through the reasons with you.

Jarvi app screenshot
Jarvi logo

Software built by recruiters for recruiters.

Get started with Jarvi and know which candidates to open first.