Rejection emails are the most reputation-sensitive message a hiring team sends. Candidates who reach the final round and lose routinely screenshot the email to LinkedIn, Reddit, and Glassdoor. ChatGPT defaults (We regret to inform you, After careful consideration, We wish you the best in your future endeavors) read as automated indifference after a six-hour interview loop. TextSight flags the six rejection-template phrases sentence by sentence, preserves your legal-safe phrasing on Light mode, and gives you a rewritten rewrite that closes the loop with dignity instead of going viral for the wrong reasons. Free to try. No card.
A candidate who interviewed four times for a senior role and then received a three-paragraph ChatGPT rejection has three places they can publish their experience by Monday morning. Glassdoor, LinkedIn, and Reddit each surface in the search results of the next candidate evaluating an offer from your company. Rejection style is one of the most-quoted categories in Glassdoor interview reviews, and it costs more in pipeline damage than the email itself takes to write.
Hiring teams write rejection emails under three pressures at once. Volume (a single loop produces five to twenty rejections), legal exposure (written feedback can resurface in a discrimination claim or a Subject Access Request), and emotional load (the recruiter is delivering disappointment to a person they spent real time with). ChatGPT lands as a relief valve for all three, and the resulting emails are uniformly templated, uniformly polite, and uniformly recognised by candidates as form letters. The realistic 2026 workflow is to use AI for the first draft, then rewrite the published version with one specific reference per candidate before it sends.
Rejected finalists screenshot rejection emails. Not all of them, but enough that any rejection sent at scale will end up on a public platform inside a quarter. A polished but generic ChatGPT rejection that gets posted does more brand damage than a delayed personal note, because the screenshot reads as institutional indifference to a candidate who spent hours with your team. The candidates posting these screenshots are usually senior, employed elsewhere, and in active hiring conversations themselves; their network is exactly the pipeline you wanted them to refer.
Glassdoor interview reviews surface above the fold for the search query company name interview, which is what candidates type before accepting an offer from your competitor. Rejection style is one of the three most common review categories alongside interview difficulty and process length. A rejection email that reads as a copy-paste job becomes a paragraph in the review, and the paragraph stays indexed for years.
Rejected finalists are the warmest possible pool for your next role at the same level. A graceful rejection keeps them open to the next requisition (which often opens within 90 days at the same company) and keeps their referral network warm. A templated rejection burns both. The math on rejection-email investment is almost always positive on a 12-month horizon, especially for senior and specialised roles where the pipeline is thin.
The LLM does not know which round mattered, which interviewer the candidate connected with, or which project from the take-home stood out. Without that signal it defaults to templates. The hiring team's job is to supply one specific reference per candidate before drafting; the AI rewriter's job is to turn the AI-assisted draft into prose that reads as written by the person who ran the loop.
A post-application rejection and a final-round rejection are not the same message. The volume, the candidate's emotional investment, and the brand-damage radius are all different, and each stage has its own template-vs-personal threshold. Treat them as four separate drafts rather than one rejection workflow.
Sent before the candidate ever spoke to a human. Volume is high (hundreds per role for senior IC openings) and the candidate's emotional investment is low because they applied alongside dozens of others. A templated note is acceptable here, but it still must not read robotic. Rewrite the opener away from We regret to inform you and keep the rest concise. Light mode handles this stage in one pass per template.
The candidate has spoken with your recruiter for 20 to 30 minutes. They invested time and they often remember the conversation. The rejection should reference one specific thing from the screen (the role scope mismatch, the seniority gap, the location constraint) rather than defaulting to a generic After careful consideration. Light mode plus one personal reference per candidate is the realistic pattern here.
The candidate has interviewed with two or more of your team members and spent hours preparing. This is the highest reputation-damage radius per candidate because the emotional investment is high, the brand surface area is wide (every interviewer has been introduced), and the candidate now feels owed a real response. Skip the templated three-paragraph structure. Reference one specific moment from an interview, name the decision factor honestly, and sign off with your real name and role.
The candidate was within striking distance of an offer. They cleared the loop and lost on a finalist comparison. These are also the candidates most likely to screenshot the rejection if it lands wrong and most likely to refer or re-apply if it lands right. The realistic pattern is a personal draft (not ChatGPT-first) with one specific strength named, the decision factor explained respectfully, and an explicit invitation to stay in touch if you mean it. Run the AI rewriter to clean the cadence, not to write the substance.
Pro at $19.99 a month standard, $14.99 a month on yearly, fits in-house hiring teams running multiple loops a month. Business at $39.99 a month standard, $29.99 a month on yearly, fits recruiting agencies and RPO teams sending rejection batches across multiple clients. Full details on the pricing page.
Billed $89.88/year — Save $30
Billed $179.88/year — Save $60
Billed $359.88/year — Save $120
Yearly billing saves 25%. View full pricing →
Rejection emails are short, emotionally loaded, and legally cautious, which makes them the easiest format for an LLM to default into. Six patterns appear in nearly every AI-drafted rejection. Candidates who have interviewed at three companies in the same week recognise the pattern instantly.
The universal opener. Appears in roughly 80 percent of AI-drafted rejections. Was already considered formulaic before ChatGPT existed; the LLM training data baked it in further. Recruiters who write five rejections a week and candidates who receive twenty a quarter can identify the opener inside half a sentence. The fix is to open with a human acknowledgement (thank you for the time you spent with us last week; appreciated the depth of your take-home submission) rather than a regret formula.
The connective phrase that signals nothing. Used to soften the decision and to fill space between the regret opener and the rejection sentence itself. Carries no actual information about the consideration. Cut it entirely; if the email already references the specific round, the project, or the strength that came up in debrief, the consideration is implicit and stated.
Thanks paragraph, decision paragraph, encouragement closer paragraph. Each is roughly the same length, each is parallel in cadence, and the structure is itself the tell. Real human rejections written by the recruiter who ran the loop have uneven lengths because the longer paragraph is the one with the specific reference. Vary the structure; let one paragraph be two sentences and another be five.
We wish you the best in your future endeavors. We have no doubt you will find the right opportunity. Best of luck on your search. ChatGPT stacks these on every rejection. They read as filler and they read as untrustworthy because the writer obviously did not have specific information to offer. Cut them. Either include one concrete future-fit comment if you mean it or sign off cleanly.
The most damaging tell. The rejection makes no reference to what the candidate actually discussed, demonstrated, or built during the loop. It could have been sent without any interviews having happened, which is what makes it land as automated. One sentence naming the round, the project, or the interviewer is enough to break the pattern entirely.
Should you wish to receive feedback, please do not hesitate to reach out. The offer is rarely taken up because it requires the candidate to ask twice and risk a second rejection. It also implicitly admits no specific feedback exists. Either provide one concrete factual sentence inside the rejection itself (legal team permitting) or skip the offer rather than dangle it.
Rejection emails carry written-record exposure. Discrimination-claim discovery, Subject Access Requests under GDPR, and pay-transparency or state-specific candidate-feedback laws all turn rejection text into evidence under the right circumstances. Treat the AI rewriter as a cadence and tone tool, not a legal-review substitute.
In US employment-law contexts, written feedback to a rejected candidate can resurface in a discrimination claim. Most employment counsel recommends keeping rejection feedback factual, role-specific, and consistent across candidates in the same loop. The AI rewriter can rewrite the opener and closer, but the substance (which decision factor you named, which round you cited) stays your judgement call. Run senior-role rejections past People or external counsel before sending.
EU and UK candidates can request a Subject Access Request that surfaces internal hiring notes alongside the rejection email. Any internal notes drafted in tone you would not want disclosed should be rewritten before they reach the candidate file. The AI rewriter is not a redaction tool; it polishes cadence on text you already intend to send.
A growing number of US states (New York City, California for certain protected categories) and EU jurisdictions require specific candidate-feedback responses on request. The realistic workflow is a templated factual response for those requests, drafted with legal review, and a separate rewritten rejection email for the candidate-experience layer. Do not conflate the two.
TextSight is not a hiring-compliance audit, an EEOC-compliance certificate, or a discrimination-risk classifier. It surfaces ChatGPT template patterns and rewrites cadence on Light, Balanced, and Maximum modes. The legal review, the policy decisions, and the accommodation language stay with your People team. Use the AI rewriter to make the message read human; use your hiring playbook to make sure it is also defensible.
A real example. A senior backend candidate cleared four rounds at a Series B fintech, lost to a stronger systems-design finalist, and was sent the ChatGPT draft below. The rewritten version is what actually went out and the candidate referred a colleague to the team three months later.
"Dear Candidate, We regret to inform you that after careful consideration we have decided to move forward with another candidate for the Senior Backend Engineer position. We sincerely appreciate the time and effort you invested in our interview process. While your background and qualifications are impressive, we have decided to pursue a candidate whose experience more closely aligns with our current needs. Should you wish to receive feedback, please do not hesitate to reach out. We wish you the best in your future endeavors. Best regards, The Hiring Team."
"Hi Rahul, thanks for the four rounds last month, the systems-design conversation with Naveen in particular was one of the strongest we have run this quarter. We went with another finalist whose recent work was closer to our payments-rails rewrite, which is the project the role will own in its first six months. That is the only delta; everything else about your loop landed well. We would genuinely like to stay in touch for the next senior backend opening we run, which is likely in Q3 once the rewrite ships. Rajeev, Engineering Manager, Payments."
The We regret to inform you opener dropped. The after careful consideration filler dropped. The vague experience more closely aligns line was replaced with the actual decision factor (payments-rails rewrite scope). The generic feedback offer dropped because one concrete factual sentence sat inside the email itself. The Best regards sign-off was replaced with the hiring manager's real name and team. Total length went from 110 to 119 words while becoming sharper. The score moved 73 points, and the candidate referred a colleague three months later.
More for hiring teams.
Rewrite ChatGPT-drafted JDs before they hit LinkedIn Jobs, Greenhouse, Lever, or Workday.
For JDs →Umbrella guide covering every business email type from outreach to follow-up to apology.
For emails →Light, Balanced, and Maximum modes for fixing flagged passages without losing voice or substance.
Read the guide →Free, Starter, Pro, Business. Yearly billing saves 25%. Hiring-manager to agency tiers.
See pricing →Free to try. No card. Pro at $14.99 a month on yearly for in-house recruiters; Business at $29.99 a month on yearly for agencies and RPO teams.