A note on reply rates before anything else. Every template below is marked with a blank where a reply rate should go. Those blanks are deliberate. These templates have not been sent at volume from this account, so there is no reply rate to report, and publishing an invented "37 percent reply rate" would be the exact behaviour this product is built to argue against.
If you send these and track the results, the numbers are yours. We will publish ours when we have them, marked as ours.
Why this kind of email works at all
Most cold email opens with a claim the recipient cannot verify and has heard before, which leaves them nothing to do but decide whether to trust a stranger.
AEO outreach has a structural advantage: the finding is checkable in thirty seconds. You are not claiming you can help. You are reporting the result of a test, and the recipient can re-run it themselves on their phone while reading your email.
That changes the job of the email. It is not persuasion, it is delivery of a fact. Which means the job is to state the fact and stop, because every extra sentence sits between the reader and the part they can check.
The structure every template shares
- The test, in one line. What you asked, where, how many ways.
- The result, as a number. Named in 0 of 8.
- Who was named instead. Two or three real competitor names. This is the line the recipient can check on their own phone.
- The reason, in one clause. Not a full diagnosis, one specific cause.
- One question. Not a meeting request, a question they can answer with one word.
No attachments in the first email. No deck. No calendar link.
Template 1: the plain finding
Subject: {Competitor} came up, you didn't
Hi {first name},
I asked ChatGPT and Gemini who the best {trade} in {city} is, eight different ways, from a logged-out session.
{Business} was not named in any of the 8 answers. {Competitor A} was named in 6, {Competitor B} in 5.
The likely reason is unexciting: your Google Business Profile lists your primary category as {wrong category}, so you are filed as {wrong thing} in most of the places these answers get built from.
Want me to send the 8 answers so you can see them?
{name}
Reply rate: ___
The last line is doing the work. It asks for permission to send evidence rather than for a meeting, and the answer is one word.
Template 2: the specific service angle
Use when the business has an obvious high-value service they are invisible for.
Subject: nobody recommends you for {service} in {city}
Hi {first name},
I ran a test on how AI assistants answer local {trade} questions in {city}. Across 8 phrasings, {Business} came up 0 times.
The one that stood out: "who does {specific service} in {city}". Three companies got named. None of them have your review count.
You have {n} reviews at {rating}. The businesses being recommended have fewer. The difference is that they say they do {service} in a place these systems actually read, and you don't.
Worth fifteen minutes?
{name}
Reply rate: ___
The review count comparison is the point of this one. A business with better numbers than the businesses being recommended is losing on something other than quality, and the email names what.
Template 3: the competitor-forward version
Riskier, because it opens with competitors rather than with the recipient. Use it only in categories where the recipient already watches those names.
Subject: {Competitor A}, {Competitor B}, {Competitor C}
Hi {first name},
Those are the three {trade}s ChatGPT recommends in {city}. I asked 8 different ways from a clean session; those three were in most of the answers and {Business} was in none.
I don't know whether that matters to you. If people are calling them instead of you, it might.
Happy to send the raw answers.
{name}
Reply rate: ___
Use real competitor names or do not use this template. The specificity is the entire mechanism, and a generic version reads as spam.
Template 4: the follow-up that is not a bump
Send four or five days after template 1 if there is no reply. Do not send "just following up".
Subject: re: {Competitor} came up, you didn't
Hi {first name},
Quick addition to my last email. I looked at why {Competitor A} keeps getting named and it is mostly two things: they are listed on {directory} and {directory}, which is where these answers get built from, and their profile spells out {n} services individually.
Both are free and take an afternoon. You can do it without me.
If you would rather someone just did it, reply and I will send what it costs.
{name}
Reply rate: ___
Giving away the fix costs you nothing here, because the fix is not the scarce thing. Doing it, consistently, for a business owner who is already working sixty hours, is the scarce thing.
Template 5: the audit-first offer
For higher value categories where a paid audit is a sensible first step.
Subject: AI visibility audit for {Business}
Hi {first name},
Short version: I asked four AI assistants who the best {trade} in {city} is, eight ways each. {Business} was named in {n} of {total} answers.
I do a fixed-price audit that covers all of it: the full question set across the assistants people actually use, every competitor that got named and their scores, the specific pages those answers are built from, and a one page list of what to change in order.
It is {price} and takes two working days. If you would rather see the free version first, say so and I will send the 8 raw answers from the test I already ran.
{name}
Reply rate: ___
If you have not settled what the audit costs, the pricing guide for agencies covers how to price it and what to put in it.
What to avoid
"AI is changing everything". They have read that sentence a hundred times this year and it predicts nothing about the email that follows.
Claiming a guaranteed outcome. Nobody controls the shortlist. Promising placement is both false and, when they discover it, fatal to the relationship. If you want proof to point to, the 60-day case study template shows how to write up real results without inventing any.
Sending the same email to a whole industry. The competitor names and the score are the parts the recipient can check. Without them you have a mail merge, and nothing in it is checkable.
Fabricating the test. If you did not run it, do not describe it. The recipient can check in thirty seconds, and the entire premise of your pitch is that you measure things honestly.
Attaching a PDF in email one. Attachments are a deliverability risk, and a PDF asks the recipient to open a file from a stranger before they have a reason to.
The volume problem
Every template above needs three facts before you can send it: the score, the competitor names, and one specific reason. Those come from an audit you ran before the prospect ever heard of you, such as the 15-minute manual method.
By hand, at the fifteen minutes that method budgets per business, twenty prospects is five hours of research before the first email goes out. The templates are not the bottleneck. The audit is.
That is what AIgnored automates: it scans a city and a niche, asks each model its own fixed set of 8 broad recommendation questions, scores every business, and marks the ones with real ratings and review counts that were never named. It does not ask service-specific questions, so a finding like the one in template 2 still needs a question you ask by hand. It also drafts the outreach from the actual scan values, so the score and the competitor names in the email are the ones the scan produced rather than numbers typed in by hand. The methodology page sets out how the matching and scoring work, including the limits, and the pricing page has a free plan for one city.
Before you send anything
Two practical rules that have nothing to do with AI.
Follow the email law that applies to you. CAN-SPAM, GDPR and PECR all have things to say about unsolicited business email, and they differ by jurisdiction. Physical address in the footer, a working unsubscribe, no misleading subject lines, and a lawful basis where one is required. This is your responsibility as the sender, not your tool vendor's.
Send from a domain you can afford to burn, and warm it. Cold outreach at volume from your primary domain is a way to lose your normal email.
Neither is optional, and both are cheaper to sort out before the first send than after.