Products › Timed Q&A
A clock changes
what you are measuring.
Written answers against a visible per-question timer
Set a limit per question and you stop measuring what somebody can research. You start measuring what they actually know, how they structure it under pressure, and how they write when there is no time to polish it.
Per-question limits you set. The clock is visible to the candidate.
assets/timed-qa-demo.mp4 — 16:9, 1920×1080, H.264. Autoplays muted on loop; keep it under roughly 8 MB and 30 seconds.Limits set per question, a round run against the clock, scored on submit.
Sixty seconds and five minutes
are different questions.
The limit is the instrument, so it should be chosen rather than defaulted. A minute asks what somebody has at their fingertips. Five asks how they organise a thought. The same question with two different clocks tests two different things.
- Per question, not per round. A factual question gets sixty seconds; a design question gets five minutes. One limit across a whole round is almost always the wrong limit for most of it.
- The candidate can see the clock. Pressure somebody knows about is a fair test. Pressure they discover at the end is a trick, and it measures surprise rather than competence.
- It auto-submits at zero, and a partial answer is still assessed — because how far somebody got in the time is itself the signal.
- It narrows the window on model-written answers. Not closes it — narrows it. Detection runs alongside, and the two together are more useful than either alone.
- Blind assessment still applies. No name, no photograph, no accent on the review screen.
- The same rubric as MCQ and conversational rounds, so a timed score sits on the same axis as an untimed one.
Round in progress — 4 of 6
limit 90s · visible countdown
Set the limits. The clock does the rest.
Five stages. The only judgement call is the one at the start, and it is the one that matters.
What the Timed round does
What the limit is actually testing+
A clock the candidate can see+
Timing, and answers a model wrote+
Scoring, with the excerpt attached+
Question sets and adaptive follow-ups+
Pipelines and evaluation+
A clock is a constraint.
Not a guarantee.
A timed round is a stronger instrument than an untimed one and a heavier obligation. Both things are true.
A human makes the call
It ranks, evidences and flags. It does not reject anyone. The pass, the fail and the pipeline stage are decisions a named person takes.
One AI gateway, no loose ends
Every model call passes through a single choke point with schema validation, caching, per-tenant rate limits and spend caps. There is no second path to a provider.
Blind is the default state
Identifying detail is absent from the assessment screen rather than hidden behind a toggle somebody can switch off on a busy afternoon.
And plainly, the part that needs saying. A timer disadvantages some people for reasons unrelated to competence — candidates writing in a second language, candidates with dyslexia or a processing difference, candidates on a slow connection or a phone keyboard. Adjusting or waiving limits where somebody needs it is your obligation, and in many jurisdictions a legal one; the software will let you set different limits, but it cannot decide when you should. A clock is also not proctoring: it narrows the window for looking something up without closing it. And we publish no accuracy figure for the scoring or the detection.
Frequently asked questions
How does a timed written Q&A round work?+
What happens if a candidate runs out of time on a question?+
How are the written answers scored, and can I check a score against the answer it came from?+
Does a timed round stop candidates pasting answers written by an AI model?+
Can I compare a timed score with scores from MCQ or conversational rounds?+
Does a timed assessment disadvantage some candidates, and can I change the limits for them?+
Published 7 September 2026 · Updated 9 September 2026 · Written by the TalbotIQ team