The Report That Shows Its Work

A hiring report is read by someone deciding whether to spend an hour of their week on an interview. If the first thing they see is three paragraphs of methodology, the report has already used up the attention it needed for the second thing.

What changed

The report now opens with a score and a skills breakdown instead of a page of caveats. Every skill links straight into the recording of the moment it was demonstrated, so a reviewer is never taking a number on faith. The link opens the replay paused on that moment instead of playing on from it. Build questions are now graded against each of their written requirements, and the report marks each one met or missed with a line of evidence. Explanatory text that used to sit on the page moved into a hover instead, there when you want it and out of the way when you do not.

The candidates table that lists everyone who has taken an assessment got the same treatment. It is now dense enough to scan the way a real hiring pipeline needs to be scanned, with status, score, AI workflow, and a recommendation all visible at once instead of spread across a wider, sparser layout. A separate "Integrity" tab, which used to sit apart from everything else on the report, was folded into the same evidence view instead, itemized per question rather than summarized as one number for the whole session.

We also found and fixed a real inconsistency. The same score could read one label in one part of the report and a different label two cards down, because the different cards on a report had quietly grown their own separate scales over time, one card using a five-way split, another using a three-way split, neither matching the scale the report's own headline number used. Now there is one scale, used everywhere on the page: Weak, Promising, Solid, or Strong. The same number means the same word no matter which card you are reading.

Why it matters

A recruiter without a technical background is trusting the report to do the judgment work for them. A report that contradicts itself, even in a small way, is a report that has to be double-checked, which defeats the purpose of writing one in the first place. Consistency is what makes the report usable by someone who cannot independently verify it against the raw session.

Quoting a candidate's weakest prompts directly under their communication score, each with its score and the area where it fell shortest, works the same way as the recording links. A score on its own asks a recruiter to trust reasoning they cannot see and can always be argued with in the abstract. Reading the actual prompt, or watching the recorded moment behind a skill, lets them decide for themselves whether the number is fair. A recruiter who still disagrees after that is disagreeing with something they can point to.

Where it points

The report's next job is cutting the clicks between "here is the number" and "here is exactly why," and giving the parts that still lean on summary text a direct link into the recording.