Finishing Fast Rarely Helps, and Your Time Can Feed a Flag

What Does the Clock Actually Record?

How long you took, and who gets to see it. Willo's knowledge base calls its time allocation a set of guard rails rather than a cut-off: you see the allocation, a timer counts up, and it runs until the answer is submitted, with any retakes counted inside it ¹. Going over does not stop you submitting.

The employer reviewing a Willo response sees exactly how long you took, with a coloured indicator for under or over ¹.

Willo also says the timer cannot be reset by restarting the assessment, closing the browser or refreshing the page, for candidates invited by email, SMS or an ATS integration ¹. So the count you see is the count the employer gets. That is one vendor documenting one product, and it holds for Willo. On other platforms a timer is a hard stop instead, and the recording ends when the allocation does, which is a different thing to plan around.

Spark Hire treats the minutes before you record as a setting of their own. That think time is set per question, anywhere from 30 seconds to 5 minutes, or unlimited, and limiting it also limits you to a single take ². The vendor tells employers to use that setting only when it is critical to evaluation, warning that candidates unhappy with a response may submit a blank answer or abandon the interview ².

A countdown before recording is planning time, not a warm-up.

Both settings were chosen from a menu before your invitation went out. That is a different fact from a rule about how the answer gets judged, and prep advice that tells you to work quickly and check your work in the same breath collapses the two.

Neither page says how a reviewer weighs the clock.

How Can the Time You Took Count Against You?

On some platforms the time you took is an input to a flag rather than a grade, and nothing in the vendor's description says a fast attempt is the suspicious one. HackerRank's knowledge base says its plagiarism model analyses code-writing patterns, time taken, copy-paste activity and tab-switching behaviour, then classifies a suspicious attempt as High or Medium ³. The feature covers coding questions only, by the vendor's own statement, and time taken is one signal in that list.

The same page publishes how often the flag is right. HackerRank puts the model's overall precision at 85%, meaning flagged sessions are correct 85% of the time, and says human oversight therefore remains crucial ³. Flagged attempts are highlighted in candidate reports, and recruiters can replay keystrokes to see the reason behind a flag ³. Candidate consent is required before a test starts when the feature is switched on ³.

That precision figure is self-reported, with no test set or recall published beside it.

What the number does tell you is that the vendor itself calls human oversight crucial. What a hiring team does with a flag is its own decision and is not on the page, so asking a recruiter what the report shows is a fair question rather than an accusation.

Stretching the clock on purpose does not follow from any of that. Working slower than you work is still not working, and a padded session hands over the same thin answer a rushed one does.

Read the argument the way the people who set the timer get it, before you treat a countdown as a judgment about you. Whether to measure speed or judgment with AI argues for letting the clock cap the exercise instead of scoring it, and for scoring the clock only in genuinely paced work, alongside a defect rate.

A cap is not a grade.

When Is Speed the Thing Being Measured?

When the platform says so, and the accommodation flow is one place it says so out loud. HireVue's 2025 explainability statement reproduces the prompt candidates see before an AI-scored interview. It offers a Request Accommodations button that relaxes the time limits for a qualified disability, with a note that the game-based section will not be presented because of its strict time requirements ⁴.

A section dropped for strict timing is a section where the clock is the measurement.

The button is a request, not a grant. HireVue says accommodation requests are directed to the relevant department of the employer managing the hiring process, and that the employer then decides whether an exemption is granted, a process the vendor places outside its own control ⁴. The wording is a sample the document says customers can alter, so the exact prompt varies. What it shows either way is that relaxing the clock changes what the assessment contains, and that is worth asking about before you accept it.

In the United States, the EEOC drew a related line in 2022, in technical assistance on algorithmic assessments that its own disclaimer says does not have the force of law. It names extended time, and an alternative version of a test compatible with accessible technology such as a screen reader, among reasonable accommodations that may be effective for some people with disabilities ⁵. Those are examples of what may work, not a menu anybody is owed.

The rule underneath that guidance carries its own exception. Under 29 CFR 1630.11, a test has to be administered so its results reflect what it claims to measure rather than a candidate's impaired sensory, manual or speaking skills ⁷. The carve-out is where those skills are the factor the test purports to measure ⁷. That is why the accommodation question keeps coming back to what the assessment claims to measure.

The guidance itself was pulled from the agency's site in January 2025, so it is read now through a web archive.

If a timer is a barrier for you because of a disability, how to ask for an accommodation on an AI interview covers the wording and where the request lands. Ask what the change removes as well as what it relaxes.

Extra time and a different assessment are not the same offer.

Use the Time You Are Given

Spend the clock on the answer instead of on beating it. Sapia's candidate explainer describes a chat interview of 5 to 7 typed questions, each answered in 50 to 150 words, and says the interview is not timed and that candidates are not penalised for how long they take ⁶. Where a format says that, finishing early buys nothing.

Sapia is describing its own product there, and a description is not an audit.

A self-inflicted loss worth avoiding is an answer cut short. A response that stops halfway because the countdown felt loud reads as thin, and thin is what the reviewer has in front of them. Nobody reading it knows you had more to say.

Two habits are a better use of the clock than speed. Name the specific thing: the project, the constraint you were under, what you decided and why. Then, where the format lets you review before submitting, read the answer back once before sending it, because the last minute is for checking rather than for handing time back.

Recruiters reading take-homes get the same instruction from the other direction. How to grade a take-home when every submission is polished argues for grading the decisions rather than the artifact, because polish costs nothing now and a clean deliverable proves little. The advice on the hiring side keeps landing in the same place: bound the task, read the reasoning behind the work, and stop treating the clock as the finding.

Outside a section built on strict timing, none of these formats says a faster answer is a better one.

See what gets scored >