This is the cold email plan we'd run for an AI SEO agency that gets companies named in ChatGPT and Perplexity and Google's AI answers. Plenty of these agencies call the work GEO. The proof takes one minute to find. Ask the AI the question their buyers ask and see who it names.
Your market is every company in your industry whose buyers ask questions before they buy. For a software agency that is anyone whose buyers type "best payroll software for small accounting firms". For a law firm agency it is any firm whose clients ask where to find a lawyer for their problem.
The signal here is unusually easy to check. Ask the question and read the answer. These are the things to look for.
Ask ChatGPT and Perplexity and Google's AI the question a buyer would ask. Note who gets named. A company that ranks on Google and never shows up in these answers has an obvious gap.
The same answer names a rival by name. That line goes straight into the email.
Sometimes the AI names the company and describes an old service or the wrong location. Owners want that fixed quickly because buyers are reading it right now.
Perplexity and Google's AI show the sources behind each answer. Open them. If the lists and articles being cited never mention the company it has little chance of being named.
Ahrefs or Semrush shows their Google rankings. A company that already ranks well has content worth building on and already understands why search matters.
Look for pages that answer buyer questions directly. Comparison pages and pricing pages and clear service pages all count. A site full of vague headlines gives an AI tool nothing to quote.
Group the list by what the answers showed.
Lead with the exact question and the names that came back.
Lead with the wrong detail. Quote the line the AI gave.
Lead with the competitor. Owners care far more about a named rival than about a new technology.
Lead with the contrast. They win the old kind of search and lose the new one.
Each sub-industry has its own questions. Build one campaign per question so the first line always matches the answer you screenshotted.
The lead magnet is the answers themselves. They are dated and screenshotted and impossible to argue with.
Ten questions their buyers ask and the answers ChatGPT and Perplexity and Google's AI give to each. List who got named and include the screenshot. Mark each answer where they are missing or described wrong.
It works because the proof is their own buyers' questions. Seeing a competitor named again and again with their own name missing does the convincing for you.
The pages those AI answers cite most often in their industry and whether the company appears on each one. It turns a fuzzy problem into a to do list.
It works because the list is the plan. Getting onto those pages is the work you do.
Rewrite their five most important pages so they answer buyer questions plainly and are easy to quote. Hand them over ready to publish.
Run the same questions again a few weeks after the pages go live and send the new answers. Make no promises on the result. The rewrite helps their human visitors either way.
It works because it is the same work the retainer pays for. The report shows the gap and the rewritten pages show how you close it.
Deliver the pages within ten days of the call.
Lead with the answer you got. A screenshot beats a claim so offer to send it.
Hi {{first_name}},
Asked ChatGPT "{{buyer_question}}" this morning.
It named {{competitor_1}} and {{competitor_2}}. {{company}} wasn't in the answer.
Ran nine more questions your buyers ask and screenshotted who came up. Want the report?
{{sender_first_name}}
Hi {{first_name}},
Perplexity describes {{company}} as "{{ai_description}}". That looks out of date.
Buyers asking about {{service}} are reading it right now.
I can show you the sources it pulled that from. Worth a look?
{{sender_first_name}}
Hi {{first_name}},
We get {{industry}} companies named when buyers ask ChatGPT and Google's AI for a recommendation.
{{proof_line}}
Want me to check where {{company}} shows up today?
{{sender_first_name}}
Swap {{proof_line}} for one real before and after of your own. If there is no result yet then leave the line out.
At small and mid-sized companies the founder or the head of marketing owns this. At bigger ones it sits with the head of search or digital.
Results depend on the offer and the list. Three outbound agencies that publish their own numbers land between one booked meeting per 1,000 people contacted and one per 2,000. One of them also tested 650,000 cold emails with a plain offer and got about 3 to 6 interested replies per 10,000 people.
Those are other people's numbers and not a promise about your market. Your market's number only comes from testing it.
We build and run this campaign for one agency per industry.
Thirty minutes on your market and how we'd reach it.