Cold Outreach Templates That Use an AI-Visibility Audit as the Hook

The best cold email you can send is one you didn't have to write from scratch: it's built around a fact the prospect can verify in ten seconds. Below are five templates built around real Litmus scan findings, ready to copy, personalize, and send.

Why generic cold email fails here

"I help businesses grow with SEO" gets deleted before the second line. Business owners get that email daily. What they don't get daily is someone naming their exact site and a specific, checkable problem with it. Every template below leads with the finding, not the offer. The ask comes last, and it's small.

Template 1: The blunt technical finding

Subject: ChatGPT can't see your menu page

Hi [Name], Ran [Business Name]'s site through an AI-visibility scanner, and it turns out [specific bot, e.g. GPTBot] is currently blocked from your menu/services page. That means when someone asks ChatGPT or Perplexity about [business type] in [city], you're invisible in the answer even if you'd normally rank well. Screenshot attached. Happy to send the fix list if useful, no charge to look. [Your name]

Template 2: The competitor-gap angle

Subject: [Competitor] is showing up in AI answers, you're not

Hi [Name], Quick one: I scanned a few [industry] sites in [city] for how visible they are to AI answer engines (ChatGPT, Perplexity, Google's AI Overviews). [Competitor]'s showing up because of [specific reason, e.g. structured FAQ content]. Yours isn't yet, mostly fixable. Attached the scan for your site. Want the 3-issue fix list? [Your name]

Template 3: The stat-led opener

Subject: A stat, then your site's score

Hi [Name], A Princeton study (KDD 2024) found that pages citing outside sources get cited by AI models up to 115% more often when they're not already top-ranked. Small, fixable content changes move the needle a lot. Ran your site through the same kind of check: your AI-visibility score came back at [X]/100, mostly dragged down by [dimension]. Full breakdown attached, free. Let me know if you want the fix. [Your name]

Template 4: The short LinkedIn DM version

Hey [Name], ran [Business]'s site through an AI-visibility audit (checks whether ChatGPT/Perplexity can actually find and cite you). A couple of quick technical fixes could change your score meaningfully. Want me to send the scan? No pitch, just the numbers.

Template 5: The follow-up after silence

Subject: Following up on the AI-visibility scan

Hi [Name], No worries if this got buried, just following up on the scan I ran for [Business Name] last week. The main issue was [one-line summary]. Takes maybe [X hours] to fix and it's the kind of thing that keeps mattering as more people search through AI tools instead of Google directly. Here's the report again if it's easier to find: [link/attachment] [Your name]

What makes these work

Every template does the same two things well: it names the business specifically and leads with one concrete, verifiable finding. What it asks for is small, a reply, not a call or a signature. Run a Litmus scan on the prospect's homepage before you write anything: the scan is what generates the specific, sourced claim these templates depend on. Without a real finding, these are just a different flavor of the same generic pitch everyone deletes.

Sequencing: don't send once and give up

Send the first email, wait 4-5 business days, send the follow-up (Template 5) with a fresh angle from the same scan (there are usually 3-5 flagged issues, so you have material for two or three touches without repeating yourself). Stop after two follow-ups. A third unsolicited email starts to read as spam regardless of how good the finding is.

Volume and targeting still matter more than wording

A perfectly worded email to the wrong prospect (a site that's already doing everything right) won't convert. Build your list first, scan every prospect, and only write to the ones with real, visible gaps. The templates are interchangeable; the specificity of the finding isn't.

Q: Should I attach the full PDF report or just a screenshot?

A: Screenshot for the first touch: lower friction, opens instantly on mobile. Save the full branded PDF for after they reply, when they're already engaged enough to read it properly.

Q: What if my finding is wrong or the site fixed it before I emailed?

A: It happens occasionally. Re-scan right before sending if the list has been sitting for more than a few days, scores change as sites get updated.

Q: Is it okay to mention I used a tool like Litmus in the email?

A: Optional. Some prospects respond well to "I use an AI-visibility auditing tool" because it sounds systematic rather than like a guess. Others don't care. Test both.

Q: What reply rate should I expect?

A: Specific, personalized findings typically outperform generic pitches by a wide margin. Expect somewhere in the 5-15% range for a well-targeted niche list, much lower for anything generic.