AI prompts for running Hunter from your AI assistant

Ask ChatGPT, Claude, or Perplexity in plain English. Every prompt returns verified contacts with sources and confidence shown inline in the chat.

Categories
79 prompts
Job-search follow-up sequence

I'm {{your_name}}, following up on job applications and interviews.

Pull my Hunter list {{list_name_or_url}} — recruiters and hiring managers I'm waiting on.

For each, pick the right note: an application follow-up (reference the role + submission date → prove fit against one requirement → deferential status ask), a post-interview thank-you (reference a specific conversation moment → reaffirm fit → mostly passive close), a competing-offer nudge where I flag {{competing_offer_deadline}}, or a graceful post-rejection "keep me in mind."

Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's job-application, interview, and candidate-side recruitment-follow-up templates: reference the specific application or interview → prove attention and fit → soft ask. Real subjects: "Noticed the [[job title]] repost: is my application still active?", "[[Job title]] panel: thank you, {{first_name}}."

Ensure the copy is real, human, and avoids any AI tell-tale signs, including em-dashes, staccato sentences, and unnatural language like: carry, shipped, ask etc.

Sales cadence follow-up sequence

I'm {{your_role}} at {{your_company}} following up on cold or warm sales threads.

Pull my Hunter list {{list_name_or_url}}.

For each, build a multi-touch cadence that de-escalates: a post-touch nurture that shares a resource, a value-tease ("reply YES and I'll send my suggestions"), an objection-reframe built around a client story, and a low-risk offer with a soft deadline — ending, if still silent, in a breakup.

One ask per email, soft and reply-oriented; only the final offer carries any urgency. Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's sales-follow-up templates (the section's largest cluster, ~35): reference prior touch → restate value/resource/proof → low-friction ask. Real lines: "Just bubbling our conversation back to the top of your inbox.", the numbered "send one number: 1 … 2 … 3 …" reply mechanic.

Ensure the copy is real, human, and avoids any AI tell-tale signs, including em-dashes, staccato sentences, and unnatural language like: carry, shipped, ask etc.

Wrong-person routing sequence

I'm {{your_role}} at {{your_company}} and I may have the wrong contact.

Pull my Hunter list {{list_name_or_url}}.

For each, keep it very short: a one-line reason for reaching out, then ask whether they're the right person and, if not, who is and the best way to reach them.

Polite, single, gatekeeper-friendly ask — no pitch. Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's sales-follow-up and seo routing templates: brief follow-up → "is it you? if not, point me to the right person." Real subjects: "Appropriate person", "Relevant person", "Better fit".

Ensure the copy is real, human, and avoids any AI tell-tale signs, including em-dashes, staccato sentences, and unnatural language like: carry, shipped, ask etc.

Breakup / permission-to-close sequence

I'm {{your_role}} at {{your_company}} and it's time to close the file on non-responders.

Pull my Hunter list {{list_name_or_url}}.

For each, draft a single graceful last touch: acknowledge the silence without guilt ("usually this means you're busy or it's not a priority"), optionally dangle one low-risk offer, and ask permission to close or hand control back ("what do you recommend as a next step?").

Leave the door genuinely open, no passive-aggression. Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's breakup templates (across no-response and sales-follow-up): concede the silence → low-risk out → control-ceding close. Real subjects: "Permission to close your file?", "Closing your file", "Goodbye {{first_name}}", "Thanks from [[company]]."

Ensure the copy is real, human, and avoids any AI tell-tale signs, including em-dashes, staccato sentences, and unnatural language like: carry, shipped, ask etc.

No-response re-engagement sequence

I'm {{your_role}} at {{your_company}} and these contacts went quiet on an earlier email.

Pull my Hunter list {{list_name_or_url}}.

For each, don't re-pitch — pivot on a single new element per touch: a different angle, a peer proof point ("[[case study company]] had the same problem"), a trigger event you noticed, a smaller ask, or a fresh insight.

Keep it low-pressure with explicit "no pressure" language, one soft opt-out-framed question each. Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's no-response (and value-nudge sales-follow-up) templates: reference the unanswered email → introduce ONE new element → low-pressure close. Real lines: "I know that's a bigger commitment, so here's a smaller version:", subject "[[Case study company]] had the same problem."

Ensure the copy is real, human, and avoids any AI tell-tale signs, including em-dashes, staccato sentences, and unnatural language like: carry, shipped, ask etc.

Webinar follow-up sequence

I ran {{webinar_name}} at {{your_company}}.

Pull my Hunter list {{list_name_or_url}} — registrants.

For each, curate rather than dump: never "here's the 45-minute recording," always the one part that matters — a 5-minute highlight clip, a single actionable takeaway, a timestamped section, or a peer's result. Serve attendees and no-shows differently ("here's what you missed").

Mostly soft or no CTA; one touch can carry a real day-specific meeting ask. Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's webinar templates: anchor to the webinar → deliver ONE piece of value → light ask. Real lines: "Instead of sending the full 45-minute recording, here's the 5-minute clip that got the strongest reaction.", subject "Clipped the best part of [[webinar name]] for you."

Ensure the copy is real, human, and avoids any AI tell-tale signs, including em-dashes, staccato sentences, and unnatural language like: carry, shipped, ask etc.

Post-event follow-up sequence

I'm {{your_role}} at {{your_company}} and I met people at {{event_name}}.

Pull my Hunter list {{list_name_or_url}} — contacts from the event.

For each, run on proof-of-attention: cite something they personally said or asked, deliver the thing you promised (an answer, a resource, a link), and make a light ask — a call, or a standing "let's compare notes," or no ask at all where staying in touch is the point. Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's event templates: reference-the-meeting → recall one specific shared moment → light ask. Real subject: "You asked about [[specific question]] at [[event name]]", and one with "No specific ask here. I just wanted to stay connected."

Ensure the copy is real, human, and avoids any AI tell-tale signs, including em-dashes, staccato sentences, and unnatural language like: carry, shipped, ask etc.

Meeting recap & next-steps sequence

I'm {{your_role}} at {{your_company}}.

Pull my Hunter list {{list_name_or_url}} — people I've just met with.

For each, send a crisp recap: the decision or direction agreed, owner/action/deadline lines for each next step, the one still-open item flagged, and a soft "let me know if I missed anything" correction check.

For internal team recaps, open "Hi team" and drop the first-name merge; for client-facing ones, personalise and close on a dated forward commitment or a check-in question. Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's meeting templates: anchor to the meeting → decisions + owner/action/deadline → surface the open item → "Let me know if I missed anything." Real line: "Wanted to capture what we discussed while it's fresh."

Ensure the copy is real, human, and avoids any AI tell-tale signs, including em-dashes, staccato sentences, and unnatural language like: carry, shipped, ask etc.

Proposal follow-up sequence

I'm {{your_role}} at {{your_company}}, chasing sent proposals.

Pull my Hunter list {{list_name_or_url}} — contacts with an open proposal.

Rather than repeat one nudge, rotate the six real angles: a plain check-in, a "do you have a decision timeline?" permission-to-plan ask, a value-add ("here's a similar project's result"), a scarcity/capacity update, a multi-thread to a secondary contact ({{secondary_contact}}), and a graceful "closing the loop" breakup.

Reference the proposal by {{project_name}} in every subject, one soft opt-in ask each. Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's business-proposal templates: reference the proposal → one reason to re-engage → soft offer of a call or help. Real lines: subject "Closing the loop on the [[project name]] proposal", "No rush on my end, but it would help me plan if I had a rough sense of when your team expects to make a call."

Ensure the copy is real, human, and avoids any AI tell-tale signs, including em-dashes, staccato sentences, and unnatural language like: carry, shipped, ask etc.

Invoice / payment-reminder sequence

I'm {{your_role}} at {{your_company}}, chasing unpaid invoices.

Pull my Hunter list {{list_name_or_url}} — contacts with an invoice outstanding.

For each, use the details in {{invoice_number_amount_due_date}} to build the escalation ladder the templates form: a plain due-date reminder, a days-overdue nudge that re-attaches the invoice, an empathetic payment-plan offer where I flag {{payment_plan_ok}}, a work-paused notice, and a final notice naming the contractual consequence.

Front-load the invoice number in every subject for thread-matching, and keep the CTA firm and singular — a closed question asking for a confirmed payment date (this is the one place follow-ups get hard, not soft). Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's invoice templates: name invoice + amount + status → reference/consequence → "Could you confirm a payment date?" Real lines: "I understand that cash flow can be unpredictable.", subject "Invoice [[invoice number]]: final notice before next steps."

Ensure the copy is real, human, and avoids any AI tell-tale signs, including em-dashes, staccato sentences, and unnatural language like: carry, shipped, ask etc.

Internship outreach

I'm {{your_name}}, a {{year_and_program}} at {{school}} with {{skill_area}}.

Pull my Hunter list {{list_name_or_url}} — companies and managers I'd intern for.

For each, ground the pitch in their real product or work, lead with a specific project idea or a most-relevant accomplishment (not vague enthusiasm), state my credentials briefly, and ask for a short conversation — with a graceful out.

Use a humble opener where I don't know if they're hiring interns. Draft a 2-touch sequence. Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's internship templates: hook (referral/trigger/humble) → credentials → specific project mapped to their work → soft conversation ask. Real line: "I've been studying [[company]]'s [[public work]] and I have a concrete idea for how I could contribute as an intern."

Cold job outreach to hiring managers

I'm {{your_name}}, a {{your_role_or_skill}} with {{years}} in {{your_domain}}.

Pull my Hunter list {{list_name_or_url}} — hiring managers at companies I'd want to work for.

For each, hook on a specific company signal (a project, initiative, funding, or recent news), mirror it with one line about a relevant accomplishment of mine, and make a de-risked, short ask — explicitly no commitment ("no open role, but a strong fit", "even an informational conversation").

Draft a 2-touch sequence, resume linked not pasted. Verify addresses, then build it in Hunter for my review.

Grounded in Hunter's job templates: company-signal hook → mirror with own accomplishment → soft 15-min ask, always de-risked. Real subject: "No open role, but a strong fit at [[company]]."

Page 1 / 7Next

Run every prompt on this page from your AI assistant.

Connect Hunter to Claude, ChatGPT, or Perplexity. Every prompt below runs on your Hunter account and returns verified contacts with provenance attached, so your assistant knows where each one came from and how confident it should be before acting.

How to use AI prompts with ChatGPT, Claude, or Perplexity?

Every prompt on this page runs inside an AI assistant with the Hunter connector installed. Once the connector is live, the assistant can call Hunter directly during the chat to find companies and contacts, verify their email addresses, draft email sequences, and pull reports on what worked.

Here's the flow:

  • Set up the connector. Add Hunter to your assistant of choice: ChatGPT, Claude, or Perplexity. The installation takes 2 minutes.
  • Fill in the tag pills before you paste. Each tag pill marks a field only you can complete: the specific number, the named customer, the exact offer. Leaving one in sends the email with the placeholder still in it.
  • Paste the prompt into your chat. The assistant reads it, calls Hunter for the data it needs, and returns the result in the same conversation. One chat replaces the CSV export, the tab hopping, and the list-rebuilding.
  • Review before anything sends. Prompts that create sequences leave the draft in Hunter, waiting for your approval before it sends. Read it and adjust before you launch.

Why does the Hunter connector matter more than a good prompt?

Without the connector, the assistant guesses, and a well-shaped prompt returns invented data with the same confidence as a real answer.

Three things go wrong:

  • The names are invented. The assistant fabricates contacts that look real. A verified email in the reply is a guess, not a Hunter lookup.
  • The list is built on guesses. Bounces show up after the send, and the sending domain takes the hit before anyone catches the pattern.
  • Reporting breaks. The next quarter starts without knowing who was contacted, what the reply rates were, or what worked well enough to repeat.

With the connector installed, the assistant runs every step through your Hunter account: Discover, Domain Search, Email Verifier, Sequences, and Reporting. Contacts come back as real Hunter records with sources and confidence shown inline, addresses are checked for deliverability before anything queues, and drafts wait in Hunter for your review.

Read more: MCPs and APIs for Lead Generation 101 Guide.

How to write outreach emails that get replies with AI?

Two-thirds of decision makers don't care whether you used AI to write your email. In a Hunter survey, recipients could only correctly identify AI-written outreach half the time. What sinks an AI-written email is the same thing that sinks any email: it reads generic. Personalize enough that the message reads like it was written to one person, not a list.

Four things AI cannot invent for you:

  • A real trigger. The message needs to reference something specific about the recipient: a project they shipped, a role they recently hired for, a piece they wrote. The assistant can research it, but you decide which trigger is worth building the email around.
  • One CTA per email. Two asks in the same email cuts your reply rate. The prompts in this library each end on a single ask. Keep that when you edit them.
  • A subject line that earns the open. Personalize with a real detail, not only the first name. Hunter's State of Email Outreach found subject lines with two custom attributes open 14 percent higher than those with one (40.2 percent vs 35.4 percent).
  • The follow-up. A single message rarely gets a reply. Hunter's State of Email Outreach shows that sending three messages instead of one lifts total replies by 106 percent (6.8 percent vs 3.3 percent), while four or more start to bring diminishing returns. Every prompt here builds sequences of at least two touches.

Want help writing that first draft? Hunter's AI Writing Assistant generates personalized outreach based on your audience and offer, so the message reads as one person writing to another rather than a template with names swapped in.

Read more: How to Write an Outreach Email That Gets Replies.

How to write AI prompts that get real results?

A good AI prompt has the same four parts every time. Skip any of them and the assistant fills the gap with a guess.

The four parts:

  • The job. State it first and state it once. "Find heads of marketing at 20 UK fintechs" is a job. "Help me with sales outreach" is a topic. The assistant needs the specific work, not the domain.
  • The inputs, marked as fields to fill in. A domain, a role, a company list, and an ICP. Leave a field for the assistant to guess and you get a guess back.
  • The output shape. Say what shape you want the answer in: a ranked list, a draft held for review, two groups returned separately. That rules out the formats you didn't ask for.
  • The stop condition. What the assistant should not do, and where it hands back to you. "Draft the sequence in Hunter but do not send anything" is a stop condition. So is "return only companies where Hunter has verified emails on file."

Every prompt in this library follows that pattern, with the fields to fill in already marked as tag pills. Copy one into your chat and read the shape before you edit.