AI evaluation and scoring

Turn every conversation into a clear hiring signal.

Zeenna evaluates candidate evidence against your configured criteria, shows the context behind each result, and prepares review-ready evidence for your team. A person decides.

  • Evidence behind every score
  • Never an auto-reject
  • A person decides

Demo: the candidate overview for Daniel Weber. His CV score rises from 74 to 91 percent once interview answers confirm two requirements. Six rated answers give an interview score of 29 out of 30, 97 percent. The overall score is the average, 94 percent, visible to the hiring team only.

Trusted by high-volume hiring teams in retail, logistics, finance and energy

EY logo
Lagardère Travel Retail logo
Geis logo
EssilorLuxottica logo
Raiffeisen Bank logo
ManpowerGroup logo
CEZ Group logo
T-Mobile logo

From answers to evidence

Evaluate consistently. Keep the full context.

Turn candidate conversations into comparable evidence, while every hiring decision stays with your team.

Each answer is rated against the benchmarks your team configured. Recruiters see the breakdown and the evidence, not a black box.

Demo: the report for Emma Clarke shows an overall score of 68 built from three criteria. Requirements are marked as met, one advisory note says the CV and the interview answers differ on retail experience for the recruiter to check, and the full transcript is available.

Candidates never see scores. Recruiters see the evidence behind each one.

The interview report

Everything your recruiter needs. Nothing decided for them.

Minutes after the interview, a finished brief is in your ATS: the score with its maths, answers word for word and the open points worth exploring in the next round.

Daniel Weber

Sales assistant · City Centre store

Completed

94 %

CV 91 %+Interview 97 %÷ 2=94 %
Qualifications and experience100 %
Soft skills and personal qualities94 %
Other questions96 %

Highlights from the interview

  • 2 years in retail
  • Weekends OK
  • Starts 1 August
Interview
14 min
Interview, report
DE → EN
Candidate questions
3

Overall score, with the maths

AI candidate summary

Neutral summary

Daniel has two years of retail experience and works mornings, afternoons and weekends. In the out-of-stock scenario he would check other stores and offer an alternative. His start date, first given as “fairly soon”, was clarified as 1 August through a follow-up. He answered all core questions.

74 %91 %

CV match, before and after the interview

  • English at workNot in the CVFrom interview answer
  • Weekend availabilityNot in the CVFrom interview answer

What the interview added to the CV

Better score than 82 % of candidates for this role.

Comparison with other candidates

Their answer

Are shifts including two weekends a month OK for you?

“Yes, I expected that.”

In their own words

“When a customer leaves happy, that makes my day.”

Answers word for word

  • Document forensics
  • Chat and CV consistency
  • Duplicate application check
  • Date consistency
0 signals to review

Advisory only. A person always decides.

Credibility check

  • Cash handling at closing time has not come up.
  • Restocking under time pressure: no example given.
No recommendation on whom to hire. The decision is yours.

Open points for the next round

2 answered1 passed to a person

Is there a staff discount?

From approved information

Questions the candidate asked

  1. 21:05 · Application
  2. 21:09 · Interview started
  3. 21:23 · Interview completed
  4. 21:24 · Report in your ATS

Illustrative timing.

Timeline

DE

Ab wann können Sie anfangen?

Start dateEN report

From 1 August

Interview in their language, report in yours

Candidate trust review

Details worth a closer look, flagged for a person.

Zeenna reviews the CV and interview for inconsistent claims, unusual document content or duplicate applications. Each signal comes with evidence you can verify with the candidate.

  • Document checks
  • Chat and CV consistency
  • Duplicate application check
  • Date consistency

Advisory only. Signals never reject anyone. A person always decides.

Demo: four advisory checks flag hidden text in a CV, a CV claim that differs from the interview, a possible duplicate application and a job that ends before it starts. Nobody is rejected; the recruiter is alerted.

Verify

CV: “Led a team of 12.” Interview: “We were three on a shift.”

People describe the same job in different words, so this is a request for clarification.

Compare and decide

Every result in one list. Sorted by score when you want it.

Completed, In progress, Rejected. Filter the list, sort by score and open any candidate to see the evidence behind the number.

portal.zeenna.ai/recruitment

Sales assistant · City Centre store

Chat interview active

Wednesday 22 July · 08:00

Rejected means the candidate declined the invitation. Zeenna never rejects anyone.

Interactive demo. Filter, sort or open any candidate.

Integrations and trust

Fits your stack. Stands up to review.

Zeenna plugs into the ATS your team already uses and comes with the evidence security, legal and procurement teams ask for.

ATS integrations

Keep your ATS. Add Zeenna to the workflow.

61supported ATS platforms

Integration depth depends on the platform and the workflow you configure. If your ATS has an API, we can talk about connecting it.

Teamio
Recruitis
SAP SuccessFactors
Personio
SmartRecruiters
Teamtailor
Onlyfy
d.vinci
Greenhouse
Workable
TalentLyft
Pinpoint
iCIMS
Lever
Oracle Taleo
Ashby
BambooHR
Jobvite
Bullhorn
Cornerstone
Homerun
Recruiterflow
Welcome to the Jungle
Zoho Recruit

Structured writeback

Evidence lands back in your ATS.

Animation: a new application triggers the configured workflow, the interview evidence is written back to the ATS candidate record, and the recruiter reviews and decides.

Security and compliance

Audited, compliant, documented.

  • ISO 27001:2022Certified
  • SOC 2 Type IICompliant
  • GDPRCompliant
  • EU AI ActReadiness
  • Grant ThorntonIndependent audit
  • AI processing stays in the EU
  • We don't train LLMs on your data

Responsible AI

Hiring decisions stay human-decided.

Zeenna structures the evidence and runs the workflow. Recruiters review, override and decide.

How we govern AI
  • Candidates always know they're talking to an AI
  • Every candidate gets the full core question set
  • Scores follow your criteria and weights
  • Integrity signals are advisory, never an auto-reject
  • A person makes every hiring decision
  • Career gaps are never scored

Evaluation questions

Transparent evidence. Human-decided outcomes.

How scoring, uncertainty, integrity signals, and recruiter control work together.

Can recruiters see why a score was given?+

Yes. The scoring experience is designed to keep the underlying responses and supporting evidence available for review.

Frequently asked questions
Can a score reject a candidate automatically?+

No. Zeenna structures the evidence and highlights the open points worth exploring in the next round, but it never recommends whom to hire. Recruiters choose their own questions, review the context and decide who moves forward.

How are integrity signals handled?+

Potential document, timeline, duplicate, or chat and CV inconsistencies are advisory. They include supporting context and never auto-reject a candidate.

Do career gaps affect the score?+

No. Career gaps are excluded from scoring.

Does every candidate answer the same questions?+

Yes. Every candidate answers the full core question set, whatever their CV says, so answers stay comparable. Follow-up questions are added on top when an answer is vague or the CV leaves something open.

Ready to see it in your workflow?

Review a Zeenna candidate profile.

See how criteria, evidence, uncertainty, transcripts, and recruiter controls appear in the workflow.

  • We set up scoring criteria for one of your roles.
  • We walk through a full candidate report.
  • We show where your recruiters review and decide.

Decorative product recap: a WhatsApp invite to Daniel, a voice answer, an interview report with an overall score of 94, an interview booked for Thursday at 10:30 and a status update synced to the ATS. Timing is illustrative.