Evidence on every line
Click a score to jump to the exact sentence that earned it, with audio. No mystery numbers.
Home / Product / AI Scorecards Stage 04 · 00:14
AI Scorecards
Every score shows its reason.
Each answer is scored against the rubric you wrote, and every score comes with a one-line reason and the part of the transcript it is based on. Recruiters can see why, agree or override in a click.
- Communication4/5
Clear, structured answer
"I let him finish and then repeated the problem back to him."
- Problem solving4/5
Specific example with an outcome
"called back two days later. He rated the call five stars."
- Role knowledge3/5
Limited CRM experience mentioned
"we used an Excel tracker, not a CRM."
English and Hindi, comfortable in bothWeighted total 3.7
1 to 5
Scale per competency, anchored to your rubric
< 5 min
From hang-up to finished scorecard
0
Candidates rejected without a recruiter
What it does
Inside AI Scorecards
Job-related criteria only
The scorer is never given age, gender, religion, caste, photo or location. Accent and grammar slips do not lower a communication score unless clarity is a stated requirement.
Recruiter override with notes
Change any score and add a note. Overrides are logged and used to check that scoring stays consistent across recruiters.
Clear review flags
Short answers, poor audio, language switches or borderline totals are marked amber for a human to listen to first.
How it works
Four steps, in order.
The same sequence runs for every candidate, so results are comparable across a whole hiring drive.
- 1
Transcribe
Speech is transcribed with speaker labels and timestamps in the interview language.
- 2
Score per rubric
Each answer is compared with your 1, 3 and 5 descriptions for that competency.
- 3
Explain
A reason and transcript excerpt are attached to every score.
- 4
Recommend
Shortlist, review or not a match, with reasons. A recruiter makes the call.
Specifications
The details.
- Competencies
- Up to 8 per role
- Weighting
- Set per competency, visible on the scorecard
- Recommendation
- Shortlist, Needs review, Not a match
- Export
- PDF scorecard, CSV, ATS sync
- Audit
- Score drift and group-level checks each month
- Override
- Any score, with note and history
Fair hiring
Rules the scorer follows, without exceptions.
Candidates are always told it is AI
The first screen names the AI interviewer, explains how answers are used and offers a human interview. Choosing a human never counts against a candidate.
Scores use only job-related criteria
Criteria come from the employer's rubric. The scorer is never given age, gender, religion, caste, photo or appearance, and accent is not scored.
AI recommends. Recruiters decide.
Every shortlist and every rejection is confirmed by a person. No candidate is rejected automatically.
Recordings deleted on schedule
Recordings are kept only as long as the employer sets, then deleted. Candidates can request a copy or deletion of their data.
Monthly consistency checks
We compare score distributions across groups candidates choose to self-report, kept apart from scoring, and flag gaps with the questions involved.
Explainable by design
Each score links to a rubric line and a transcript excerpt, so any decision can be explained to a candidate or an auditor.
Questions
AI Scorecards questions
Is the AI the decision-maker?
No. It recommends. A recruiter approves every shortlist and every rejection.
How do you check for bias?
Scores are compared across groups that candidates choose to self-report, separately from scoring. Gaps are reported to your admin with the questions involved.
Can candidates see their scores?
You choose. Many teams share a short feedback note; scores stay internal by default.
Free trial25interviews
New role · ready to invite
Your next 25 applicants
are one link away.
lancdon.com/i/your-first-roleRun 25 interviews free
No card needed · Set up in an afternoon · Or connect with our team