88% of Employers Claim a Rubric. Write a Follow-Up the Debrief Can Paste

Your last interview was most likely transcribed, tagged and attached to a scorecard, and the people who decide whether you move forward will meet without you. Matt Charney, Principal Analyst at Kyle & Co and author of an October 2026 report with Textio, puts the problem plainly: "Eighty-five percent of employers already record interviews, and candidates still can't tell you how they were judged." Your follow-up note lands in that gap, and right now most notes do nothing there.

What did the Kyle & Co and Textio report find?

Employers have built structured, tool-assisted interviews and rate them highly. Candidates rate the same interviews close to zero. The report used mirrored surveys of 257 employers and 174 job seekers, fielded in August 2026, and The Globe and Mail covered it on Oct 19.

The main findings:

• 85% of employers use interview intelligence tools, and 97% are at least piloting one.
• 88% say interviewers always or usually ask the same core questions and follow a defined scoring system.
• 48% of candidates said their most recent process felt at least partly improvised.
• Employers gave their own process a Net Promoter Score of +48.25. Candidates gave the process they had just been through +2.87. Business Wire described this as job seekers being 8x more likely than talent leaders to rate an interview process poorly.
• Only 14% of employers have reached the maturity level where someone analyzes the interview data or compares it with how hires later perform.

Some caveats. Lavalier, Textio's own interview intelligence platform, sponsored the report. The respondents were nearly all in the U.S., and one newsletter noted that Textio sells the category of software it measured. The word "record" also isn't the survey's wording: the formal statistic is use of interview intelligence tools, and "record" comes from Charney's quote. His next line is the useful one: "That's the difference between a recording and a record."

How does a hiring committee debrief decide candidates?

A debrief is the meeting where interviewers compare scorecards and argue for or against a candidate. It decides based on what was written down, not on what people remember. Each interviewer submits ratings and comments, the tool may attach transcript excerpts, and the group goes through the competencies one at a time.

That's where the recording actually gets used. Greenhouse's 2026 guide describes interview intelligence platforms as connecting interview evidence to scorecards, structured evaluation criteria and ATS workflows, and companies like BrightHire build them. In practice, what you said becomes transcript snippets tagged against competencies. But only 14% of employers analyze that data systematically, so it isn't feeding some model that judges you over months. It gets read at one point: the debrief, when someone opens your scorecard and looks for evidence. Write your follow-up for that point.

What gets written down is lopsided. Textio's 2025 study of 10,377 assessments and its accompanying report found:

• 39% more feedback is documented when a candidate is not getting an offer.
• Candidates who get offers are 12x more likely to be described as having a "great personality."
• 84% of rejected candidates never received feedback.

Textio co-founder Kieran Snyder has pointed to this as likability bias. The effect in the room is simple. The interviewer who wants to reject you shows up with paragraphs. The interviewer who wants to hire you shows up with "great energy, would enjoy working with." When those two disagree in a debrief, the detailed critique usually beats the vibe.

So your follow-up has a clear job. Give the interviewer who's already on your side evidence about your skills, in a form they can quote. They don't need more proof that you're pleasant.

Does a thank-you email actually get you hired?

Sometimes, and mostly as a tiebreaker. In TopResume's survey, 68% of hiring managers and recruiters said a thank-you note affects their decision. In Robert Half's survey, only 27% said a thank-you could tip the scales when candidates had equal skills and experience.

Those numbers don't contradict each other. A note "affecting" a decision usually means it signals interest or confirms professionalism. Tipping a split decision takes something a scorecard field can hold, and gratitude doesn't fit in a field called "Systems design: evidence."

A thank-you email that actually gets you hired is the right-hand column with one sentence of thanks at the top.

How do you find the structured interview scoring rubric?

Work it out from the posting and the questions you were asked, or ask the recruiter. Don't assume anyone explained it: the report's first free recommendation is that interviewers explain the evaluation criteria out loud at the start, which tells you most don't.

A structured interview scoring rubric is a fixed list of competencies, each with a rating scale, that every interviewer on the loop scores against. 88% of employers say they have one. 48% of candidates felt the process was partly improvised. Both can be true: the rubric exists in the ATS, and the interviewer glanced at it five minutes before the call.

To rebuild it:

1. Take the competency nouns from the posting. "Owns data quality end to end" means a data quality or ownership competency. Ignore the perks and the mission paragraph.
2. List every question you were asked and label what each one was probing. "Tell me about a time a launch slipped" is about execution or stakeholder management, not storytelling.
3. Find the overlap. A competency that shows up in both the posting and the questions is almost certainly a rubric line.
4. Ask the recruiter directly. "Which competencies was today's panel scoring?" is a reasonable question, and the worst answer is no answer.
5. Keep the three or four strongest. Those become the section headers of your note.

Steps 1 and 3 are the analysis Refolk runs when it scores how well you fit a posting: it breaks the posting into requirements and shows which ones your own history covers. Run that before the interview and you walk in with the likely rubric already drafted. Afterward you only add the questions you were actually asked.

Who is actually trained to score you?

Probably not the person holding the scorecard. Rubric scoring is rarely a credential anyone lists. In Refolk's index of professional profiles, only 48 U.S. profiles list "Structured Interviewing" as a skill, and only 4 in the UK. The U.S. matches include a VP of Talent Acquisition, a technical recruiter and a Recruiter III, at employers such as Apple and Oracle (OCI).

Compare that with who is building the tools. In the same index, 151 profiles put "interview intelligence" in their headline. They are mostly founders and CEOs, at companies including BrightHire, Honeit, InCruiter and Exterview AI, and they cluster in the San Francisco Bay Area and Bengaluru.

Treat these counts as directional, not a census. They are self-tagged and noisy, and the U.S. results include at least one librarian. The direction still holds: more people build rubric software than list rubric scoring as a skill. Interviewers are often filling in a rubric they never trained on, between meetings, from memory, with a transcript they don't have time to reread.

That's why mapping the evidence for them works. If your note already has the structure the scorecard wants (competency, what you said, proof), the interviewer can copy it instead of reconstructing it.

What to write in a follow-up email after an interview

Write one short block for each competency the panel scored: the label, the point you made in the room, and one proof point. Use the posting's wording for each label, because the 88% of employers who claim a defined scoring system are scoring against named criteria. Add a correction if you fumbled an answer, keep the whole thing to one screen, and send it the same day, while scorecards are still being written.

The structure:

1. One line of thanks. Name the role and the people. That's all the courtesy you need.
2. Three or four competency blocks. Use the employer's own wording from the posting as each header, so the label matches the scorecard field.
3. Under each header, two sentences. First, what you said, in the words you used. Second, the proof: a number, a named system, a link.
4. One correction or addition, if needed. Better to fix it yourself than leave it in the transcript.
5. No new asks. Don't request feedback, a timeline or a second call in this message.

A skeleton you can fill in:

Subject: Following up on today's interview: [Role]

Hi [Recruiter], thank you and [Interviewer names] for the time today. A few points from the conversation, in case they're useful to the panel:

[Competency 1, posting's wording]. I described [the project, as you said it]. The result was [the exact figure you gave in the room], and [link or artifact] shows the before and after.

[Competency 2]. When [Interviewer] asked about [question], I talked about [situation]. The part I'd underline is [the decision you made and why].

One correction. I said [X] about [topic]. The accurate figure is [Y].

Thanks again, [Name]

Send it to the recruiter, and to each interviewer if you have their addresses. The recruiter is the person most likely to get it in front of the panel before scorecards close.

Writing this from scratch after every loop is tedious, and it's where people slip back into "great conversation, excited about the mission." Refolk drafts it from the same history your resume came from, so the competency blocks start with your real projects and numbers. You edit, you don't start from a blank page.

Why the transcript punishes embellishment

With 85% of employers running interview intelligence tools, assume a verbatim transcript exists, and any figure in your follow-up that's bigger than the one you said out loud is a contradiction anyone can check. "Cut costs by about a third" in the room and "cut costs 45%" in the email is a discrepancy a skeptical interviewer can find in seconds.

The fraud question raises the stakes. Employers often lack clear rules separating preparation from inappropriate live assistance, and candidates describe being wrongly accused of fraud. A note that reads as generic and machine-polished, or that drifts from what's in the transcript, fits the pattern a suspicious reviewer is looking for.

Rules that keep you safe:

• Reuse your exact figures. If you said "roughly," write "roughly."
• Correct out loud. "I said X; the accurate figure is Y" reads as honest. Quietly changing the number reads as the opposite.
• Write in your own voice. The panel has heard you talk. A note in a different register stands out.
• Match your resume. The number in the note, on the resume and in the transcript should be the same number.

The last rule is where most candidates trip, because their resume was written months ago with different rounding. Refolk writes the resume from your own history and tailors it per posting from that same source, so the resume, the cover letter and the follow-up all draw on one set of facts instead of three versions you have to keep in your head.

FAQ

Are job interviews recorded without the candidate being told?

The Kyle & Co and Textio report measures whether employers use interview intelligence tools (85% do). It doesn't measure how they disclose it. The practical move is to ask when you schedule: "Will this interview be recorded or transcribed?" It's a normal question, and either way you should treat everything you say as quotable.

Should I still send a regular thank-you email?

Yes, but fold it into the debrief-ready note instead of sending it separately. One line of thanks covers the 68% of hiring managers and recruiters who say a note affects their decision. The competency blocks underneath cover the split decisions, where Robert Half found courtesy alone tips the outcome only for the 27% who see candidates as equal.

How long should the follow-up be?

One screen. Each competency block is two sentences: what you said, then the proof. The interviewer reading it is filling out a scorecard between meetings. A note they can copy from in a few seconds gets used, and a long essay gets skimmed and archived.