AI STARTUP · LOCAL SEO + GEO/AEO · CLIENT UNDER NDA
From a blank Google listing to the number one answer in ChatGPT.
A brand-new AI startup with no search presence of any kind. In about four months it held the top three of the Map Pack. Not long after, it was the first name ChatGPT, Perplexity, Gemini and Claude gave when asked for the best product in its category.
Key results
Starting from nothing
Most SEO engagements start with something. A domain with some age, a Google Business Profile with a few reviews, a handful of pages that rank for the brand name. This one started with none of it. A new company, a new product category that buyers did not yet have a name for, and a founding team that needed customers before the runway ran out.
The harder problem was the one nobody was talking about yet. The buyers for this product were not only searching Google. They were asking ChatGPT and Perplexity which tool to use, and those answers named competitors. Being invisible in AI search was going to cost more than being invisible in Google, and there was no playbook for fixing it.
What we actually did
In the order we did it. The sequence matters more than any single step.
Built the Google Business Profile like it was the product page
Most profiles are an afterthought. We treated this one as the primary asset: the most specific primary category Google offered, every relevant service listed, a description written for the searcher rather than the founder, and a photo cadence that kept the listing looking alive. Review generation ran as a system from week one, following Google policy to the letter, because a listing that earns reviews steadily outranks one that collects them in bursts.
Wrote the pages that defined the category
When a category is new, the first clear explanation of it tends to become the reference. We wrote that explanation: what the product does, who it is for, how it compares to the older way of doing things, and what it costs. Plain language, structured so a person could read it and a model could quote it. These pages became the source that AI assistants pulled from when users asked about the category.
Made the site legible to machines, not only to people
Structured data for the organization, the product, and every question a buyer might ask, so that what the pages said in prose was also said in a format a crawler could not misread. Consistent naming across the site, the profile, and every third-party listing, so that a model encountering the brand in five places saw one entity rather than five.
Earned mentions in the places models read
AI assistants do not only read your site. They read the directories, the comparison articles, the industry roundups, and the discussions that cite you. We got the company listed, reviewed, and referenced in the sources that matter for its category, because a model that sees a name repeated across trusted sources treats it as the answer.
Measured AI visibility the way we measure rankings
We tracked, every week, what ChatGPT, Perplexity, Gemini and Claude said when asked the buyer's questions. Not once, because the answers drift. That tracking is how we knew when the company moved from absent, to mentioned, to first, and it is how we caught and fixed the cases where a model described the product wrong.
What it took, honestly
About four months of consistent, structured work to reach the top three of the Map Pack. Longer than that for the AI search result, because that depends on sources outside the site updating, and those move on their own schedule. There was no trick and no hack. The work compounds, which means the early months look slow and the later months look sudden.
The result that mattered most was not a ranking. It was that enterprises, including Fortune 500 companies, started reaching out on their own, having found the company by asking a question and getting its name as the answer.
Does this apply to you?
If your buyers ask an AI assistant for a recommendation before they ever search Google, and the answer names a competitor, yes. That is true for more categories every month, and it is true for local services, software, and professional firms alike.
If your market is purely local and your buyers still live in the Map Pack, the first half of this playbook applies and the second half can wait. We will tell you which one you are.
About Mining Wells
We're on a mission to fix bad marketing.
Maybe:
- You are spending thousands on marketing tools, ads, and your website, with zero revenue increase to show for it.
- Every campaign you have tried gets minimal results.
- You have a great product that nobody seems to find.
- You are getting interest, but it never converts to a sale.
- You have a low retention rate.
- You have been paying a marketing agency for over a year and have not seen results.
You are not alone. Many founders and leaders live with the results of bad marketing without ever finding the reason.
And often that is because it can be many reasons. Sometimes it is the wrong ICP, sometimes the wrong messaging, sometimes the wrong targeting chasing impressions.
We are here to take the hard guesswork out and provide that clarity before it is too late.
At Mining Wells, we help founders and leaders grow their businesses the right way.
Tired of bad marketing?