Send clients candidates who hold up
Every new client brings a new role to screen for. TalentDecrypt turns the job description into a proctored test, and gives you a scorecard you can forward with the evidence attached.
Tests / Backend Engineer, hiring assessment / Sneha Menon
Sneha Menon
Backend Engineer
Finished
OVERALL SCORE
0 / 100
Strong result
PERCENTILE
0th
Rank 13 of 214 candidates
PROCTORING COVERAGE
SESSION
- Invited
- 12 Sep, 10:04 am
- Started
- 14 Sep, 7:30 pm
- Finished
- 14 Sep, 8:34 pm
OVERALL
0/100
+26 vs cohort
ACCURACY
0%
11 of 12 correct
TIME USED
0/75 min
Finished 11 min early
INTEGRITY
Clean
0 flags, 3 feeds
Skill analysis
Sneha Cohort average
The scorecard you can forward to a client · sample data
What a test for this role can look like
One example. You choose the sections, the question types, the length and how closely it is watched.
- Draft it from a job description, or build it by hand
- Review every question before a candidate sees it
- Reuse questions from your own bank
Customer Support Associate — client screening
Voice and chat support, international process
- Duration
- 45min
- Sections
- 5
- Questions
- 24
- 01×2
Spoken response to a customer complaint
Spoken answer - 02×3
Listen to a call, then answer
Listen, then answer - 03×2
Read a script aloud
Read aloud - 04×2
Written reply to an escalation email
Written, marked by AI - 05×15
Process and product scenarios
Multiple choice
Proctoring: Webcam + screen + phone camera
Why agency screening is hard to get right
- 01
A new role to screen for every week
One client needs Java developers, the next needs support agents, the next needs verification engineers. Building a fresh test by hand for each requirement takes time your recruiters do not have.
- 02
Your name is on every submission
When a candidate you presented fails the client’s first technical round, the client questions your screening. You need proof of skill before the profile leaves your desk.
- 03
Proxy test-takers and impersonation
Remote screening at volume attracts people who take tests for others. A webcam alone does not show a second person in the room or a second device on the desk.
- 04
Spoken English decides support and BPO roles
A recruiter calling each applicant to judge communication does not scale, and two recruiters rarely score the same call the same way.
From a client requirement to a shortlist
- 1
Turn the client’s job description into a test
Drop in the client’s job description. You approve the skills, the shape of the test and every question before it goes out.
- 2
Invite the whole pipeline
Send invites in bulk. Candidates take the test in the browser. For high-stakes roles, turn on CheqMate AI — and TriNetra, which makes the candidate’s own phone a second camera on the desk.
- 3
Screen communication in the same sitting
Add SpeakPro spoken questions for support, sales and client-facing roles. Each answer is recorded, transcribed and scored by dimension.
- 4
Shortlist from one scorecard
One scorecard per candidate: skill scores, the integrity record and an AI summary. It decides who you put forward, and it is the evidence behind the submission.
What you would use
- AI Test GenerationA new client requirement becomes a reviewed test without pulling a subject expert off billable work.
- CheqMate AICheqMate AI watches the screen and the device; TriNetra sees the desk and the room, which is where a proxy or a second phone shows up.
- SpeakProSpoken-English screening for BPO, support and sales roles is scored the same way for every applicant.
- Candidate ScorecardsOne scorecard per candidate, every score linked to the answer behind it, so the profile you submit holds up.
Roles you can assess
- Java developer
- Full-stack developer
- Data analyst
- QA engineer
- DevOps engineer
- Customer support associate
- Technical support engineer
- Inside sales executive
- Verification engineer
- Embedded engineer
- Finance and accounts associate
- Contract IT staff
Common questions
Can we run separate tests for each client?
Yes. Each test has its own questions, settings and invite list. Questions you want to reuse go into your organisation’s question library, tagged by topic, so the next test for a similar role starts from what you already have.
How quickly can we build a test for a new requirement?
AI test generation drafts a test from the job description. You confirm the skills, then the blueprint, then review the questions before any candidate sees them. You can regenerate a single question with a hint if it misses the mark.
What stops someone else taking the test for the candidate?
Someone checks the candidate’s ID and their room on a live call before the test starts. During it, CheqMate AI reads the screen, the webcam and the device, and TriNetra turns the candidate’s own phone into a second camera on the desk. Every flag is recorded against the question they were on, with the evidence attached.
Can we connect TalentDecrypt to our own systems?
Yes — invites can be created and results read through an API, so candidates and outcomes flow into the system your recruiters already live in.
See it with one of your own roles
Bring a job description to the demo. We will build the test with you and show you the report a candidate produces.