Verified against ChatGPT · 2026-08-09
Write a recruiter outreach message that references something the candidate actually did
Drafts a short, non-generic outreach message to a passive candidate anchored in a specific, real piece of their public work, built to survive the candidate's built-in skepticism toward mass recruiter spam.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Write a short outreach message to a passive candidate who is not job-hunting and gets recruiter messages constantly. The only way this message gets a reply is if it doesn't read like every other one they ignore. CANDIDATE AND WHAT THEY DID THAT CAUGHT YOUR ATTENTION Gave a conference talk last month on migrating a payments system off a legacy monolith with zero downtime. ROLE YOU'RE REACHING OUT ABOUT Staff Backend Engineer leading a similar legacy-to-microservices migration for a fintech scaleup. WHY THIS PERSON SPECIFICALLY, NOT JUST "THIS TITLE" We're six months into the exact kind of zero-downtime migration she described, and are missing someone who's actually done it before. WHAT YOU CAN ACTUALLY OFFER THAT'S DIFFERENT FROM WHERE THEY ARE She'd own the migration architecture directly instead of being one of six engineers executing someone else's plan. HOW YOU FOUND THEM Found the talk recording on the conference's YouTube channel. Open with the specific thing they did — not "I came across your profile," which every ignored message also says, but the actual project, talk, post, or shipped thing, named precisely enough that a screenshot of it would match. Connect that specific thing to the role in one sentence, not a paragraph — the connection should be a single clear reason, not a list of ways they're "a great fit." State plainly how you found them; candidates are more likely to respond when the sourcing method isn't mysterious, and pretending a cold LinkedIn search was an organic discovery reads as manipulative once they notice, which they usually do. Keep the whole message under 120 words. Do not include a bulleted list of the role's responsibilities or a link to a full job description in the first message — that's a second-message asset, not an opener. Do not use the word "exciting," "passionate," or "rockstar," and do not open with the recipient's first name followed immediately by flattery about their profile in general terms, since that pattern is the single most-ignored recruiter opener that exists. End with a low-friction ask — a single question that's easy to answer with one line, not "let me know if you'd be interested in learning more," which asks the candidate to do the work of figuring out how to respond. OUTPUT FORMAT 1. Subject line (if email) or opening line (if InMail/LinkedIn), under 10 words. 2. The message body, under 120 words. 3. One alternate opening line in case the first doesn't land, using a different specific hook if more than one was given.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
Passive-candidate outreach fails at the volume most recruiters send it, and the reason is structural rather than stylistic: recipients pattern-match on a handful of tells (generic profile flattery, vague "exciting opportunity" language, a bulleted responsibilities list in message one) within the first two lines and stop reading, regardless of how good the actual role is. Explicitly naming the words and patterns to avoid works because GPT-5.1's default register for "recruiter outreach" leans exactly toward that overused vocabulary, since it's overrepresented in the training data for the genre — a negative instruction list is doing real work here, not padding, because without it the model reaches for "passionate," "rockstar," and generic profile praise almost by default. Requiring the hook to be specific enough that "a screenshot would match" forces concreteness that a vague instruction like "personalize the outreach" doesn't reliably produce, since a model can satisfy "personalize" with a paraphrased generality that still reads as templated. Stating the sourcing method plainly rather than let the model default to a mysterious "came across your profile" phrasing matters because that phrase is itself a well-known tell that experienced candidates recognize as evasive, and transparency about sourcing measurably improves response rates because it removes one layer of suspicion the reader would otherwise have to resolve before deciding whether to trust the rest of the message. The word-count ceiling forces every other instruction to compress into what actually matters, since without it the model will default to including the full role rationale in message one, which the low-friction-ask instruction is specifically designed to prevent by pushing detail into a second message the candidate has to opt into first.
What you get back
Subject: Zero-downtime migration talk. Body: Hi Priya — watched your talk on the payments-system migration off the legacy monolith with zero downtime. We're six months into almost the identical migration at a fintech scaleup and don't have anyone on the team who's actually done this before, not just read about it. This role would have you owning the migration architecture directly rather than executing someone else's plan. Worth a 15-minute call to compare notes, even if it's not the right timing? Found your talk on the conference's YouTube channel, for full transparency.
Verified against
ChatGPT GPT-5.1 · 2026-08-09
Changelog
- 2026-08-09 — 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
