HR & Management

Verified against ChatGPT · 2026-08-10

Write a final-round rejection email that gives real feedback without creating legal exposure

Drafts a rejection email for a candidate who made it to the final round, giving one specific, useful reason without comparative statements or promises the company can't stand behind.

ChatGPT (GPT-5.1)4 fillable variables

The prompt

Ready to copy — highlighted parts are example details you can swap.

You are writing a rejection email for a candidate who made it to the final round of interviews — further than a generic "thank you for applying" rejection, and deserving of a more specific response, but one that stays within safe, defensible bounds.

CANDIDATE AND ROLE
Marcus, final-round candidate for Senior Product Manager, Growth team.

ACTUAL REASON FOR THE DECISION
The role needs someone who has run paid-acquisition experiments hands-on in the last two years; Marcus's experience there was managerial oversight rather than direct execution.

STRENGTHS WORTH NAMING
His breakdown of our onboarding funnel in the case study interview was sharper than any other candidate's — he found a drop-off point our own team had missed.

FUTURE RELATIONSHIP
We'd genuinely like to keep him in mind if a more senior, strategy-focused role opens in the next two quarters, and our VP of Product will personally reach out if that happens.

Write the email to name one specific, defensible reason for the decision — tied to the role's actual requirements, never a comparison to the other finalist ("we chose someone with more X" invites the reader to wonder what X they supposedly lack, which is both unhelpful and legally sloppier than describing what the role needed). State the reason as a gap against the role's needs, not a judgment about the person overall — "the role needed deeper hands-on experience with Y in the first 90 days" is defensible; "you're not senior enough" is not.

Name at least one genuine strength specifically, using something they actually demonstrated in the process, not generic praise — this should read as evidence you paid attention, not as a consolation prize.

Do not promise future consideration, a specific other role, or that the company will "keep them in mind" unless that is genuinely true and someone is actually going to follow up — an insincere promise here is worse than no promise, since it sets an expectation that will go unmet and damages the company's reputation with a candidate who was otherwise a near-miss.

Do not include any statement that could be read as feedback on protected characteristics, personality judgments, or comparisons to specific other candidates. Keep the tone warm but brief — a final-round candidate deserves respect, not an essay.

OUTPUT FORMAT
1. Subject line.
2. Email body, three short paragraphs: the decision stated plainly and early, one specific defensible reason plus one genuine strength, and a warm closing that matches the future-relationship input exactly (no embellishment beyond what was stated as true).

Customize

Optional — swap in your own details for the highlighted parts above.

Why this works

The instruction to avoid comparative language ("we chose someone with more X") is doing the most legally and practically important work in this prompt: comparative rejection language invites the candidate to reverse-engineer what they supposedly lack relative to another specific person, which both feels worse to receive and creates a paper trail that reads badly out of context if the decision is ever scrutinized, whereas a gap-against-role-requirements framing is defensible because it's about the job's needs, not a ranking of people. GPT-5.1 defaults toward warm, softening language when asked to write a rejection, which without guardrails tends to produce vague reassurance ("you were a strong candidate, it was a tough decision") that says nothing specific — requiring one named, role-tied reason forces the output past that vagueness into something a candidate can actually learn from, which is the entire point of giving final-round feedback at all instead of a form rejection. The instruction against insincere future-consideration promises exists because a model asked to write something "warm" will reach for exactly that kind of soft closer by default, and here it's explicitly gated behind whether it's genuinely true, which requires the requester to have actually decided that before generating the email rather than letting boilerplate warmth manufacture a commitment nobody intends to keep. The prohibition on personality judgments and protected-characteristic-adjacent language matters because "not senior enough" or similar phrasing, even when not legally protected-characteristic language itself, reads as a judgment about the person rather than a fit assessment against the role, and that distinction is exactly the line between defensible hiring communication and something that sounds like a pretext if ever examined.

What you get back

Subject: Update on the Senior PM, Growth role. Body: Hi Marcus — after a lot of deliberation, we've decided to move forward with another candidate for the Senior Product Manager, Growth role. The role specifically needs someone who has run paid-acquisition experiments hands-on in the last two years, and your recent experience has been more on the managerial-oversight side of that work. That said, your breakdown of our onboarding funnel in the case study interview was the sharpest of any candidate we saw — you caught a drop-off point our own team had missed. If a more senior, strategy-focused role opens on our team in the next two quarters, our VP of Product will personally reach out.

Verified against

ChatGPT GPT-5.1 · 2026-08-10

Changelog

  • 2026-08-10 Initial publish, verified against ChatGPT GPT-5.1.

Need this built into your business?

If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.

EXPLORE WHAT SCULT BUILDS
Write a job description that matches the actual hiring bar instead of a wish list nobody meetsTurns a hiring manager's notes into a job description built around a true must-have vs. nice-to-have split, so the posting attracts candidates who can actually clear the bar instead of scaring off good ones with an inflated requirements list.ChatGPT (GPT-5.1)2026-08-08Build a candidate scorecard that stops a 4-person panel from scoring the same interview four different waysGenerates a per-candidate scorecard tied to the specific competencies this role was actually interviewed for, so a hiring panel's individual scores are comparable instead of reflecting how generous or strict each interviewer happens to be.ChatGPT (GPT-5.1)2026-08-09Calibrate what a 1 through 4 actually means before four interviewers disagree about it in the debriefProduces the scoring rubric itself — the shared definition of what each competency and each score level means — that a hiring panel reviews together before interviews start, so scorecards filled out later are measuring the same thing.ChatGPT (GPT-5.1)2026-08-10Generate interview questions with the follow-up probes that catch a rehearsed answerWrites a set of role-specific interview questions, each paired with two or three follow-up probes designed to surface real depth or expose a memorized answer, plus a flag on any question that risks drifting into illegal or discriminatory territory.ChatGPT (GPT-5.1)2026-08-11
All HR & Management prompts

Check your AI visibility

One URL in, a 0–100 score and the exact fixes out.

RUN THE CHECK

Browse all the tools

15 tools across six categories
13 of them never send your data anywhere

Free · No signup · No trial clock

SEE THE DIRECTORY