.jpg)
Search is changing.
A customer looking for a dentist, architect, plumber, attorney, interior designer, or marketing agency in Portland can still turn to Google. But increasingly, they can also ask ChatGPT, Gemini, Perplexity, or another AI tool a much simpler question:
“Who should I hire?”
That creates a new problem for businesses.
Ranking well on Google does not automatically mean an AI system will recommend you. And unlike a traditional search result, an AI recommendation can pull together information from your website, reviews, directories, publications, business profiles, and other sources across the web.
So we wanted to understand what the Portland businesses showing up in these recommendations actually have in common.
At Belorado GEO, we analyzed AI visibility across six commercially meaningful Portland industries: architecture, interior design, personal injury law, dentistry, plumbing, and SEO/web agencies.
What we found was less about one secret “GEO ranking factor” and more about something broader:
The businesses with strong AI visibility tend to have multiple independent parts of the web agreeing about who they are, where they operate, what they do, and why they're credible.
We built an exploratory matrix of 60 commercial-intent prompts, with 10 searches across each of six industries.
The prompts included broad recommendation searches, service-specific searches, reputation and trust searches, and more specialized queries.
We then investigated the businesses that repeatedly surfaced, looking for patterns across:
There is an important limitation to the recommendation-frequency portion of this research.
The exploratory prompt test was conducted in an environment that contained previous Portland business context. Because of that, we do not consider its recommendation counts a clean independent benchmark, and we're not presenting those percentages as proof of how ChatGPT ranks Portland businesses.
Instead, we used the recommendations to identify businesses for deeper signal research and focused our conclusions on patterns that could be independently corroborated.
And those patterns were remarkably consistent.
One of the clearest patterns wasn't on the businesses' websites at all.
It was everywhere else.
Businesses with strong visibility repeatedly appeared across independent sources that corroborated the same information about them.
Architecture firms appeared through professional organizations, architecture publications, awards databases, industry profiles, and local publications.
Interior designers appeared across platforms like Houzz, local editorial coverage, awards, business directories, and review platforms.
Law firms appeared in legal directories, professional profiles, review datasets, and publications.
SEO agencies appeared in agency directories, comparison articles, industry lists, case studies, and other independent sources.
In other words, these companies weren't solely telling the internet that they were good at something.
Other parts of the internet were saying it too.
That makes third-party corroboration one of the strongest correlations we observed with AI recommendation visibility.
Reviews were extremely prominent in high-frequency local consumer categories.
Portland plumbing and dentistry provided some of the clearest examples.
Frequently visible businesses often had substantial Google review footprints, complete local profiles, clear service information, and strong Maps visibility.
That makes intuitive sense. For a consumer trying to choose a plumber or dentist, reviews provide an enormous amount of evidence about whether a business exists, serves the area, performs a particular service, and has satisfied customers.
But architecture showed us why reviews aren't the whole story.
ZGF Architects' Portland office had only 21 Google reviews in the data we examined. Scott Edwards Architecture had 14. Propel Studio Architecture had six.
Yet these firms had substantial evidence elsewhere.
ZGF, for example, has major industry recognition, awards, documented projects, professional authority, and extensive content establishing what the firm does and its connection to Portland.
That leads to an important distinction:
AI visibility does not appear to simply reward the business with the most Google reviews.
In consumer services, reviews and local reputation appear especially important. In reputation-driven professional services, recognized expertise, documented projects, industry authority, and third-party validation may compensate for a relatively small review profile.
Simply having a website wasn't the pattern.
Having a website that leaves very little ambiguity about the business was.
The recurring businesses tended to explicitly connect:
Business → Location → Service → Specialty → Proof
A strong dental website, for example, doesn't merely say it's a dentist. It identifies its Portland location, individual doctors, hours, implants, emergency dentistry, whitening, sedation, and other specific services.
A strong plumbing page doesn't just claim to offer plumbing. It establishes the location served, specific plumbing services, emergency availability, common problems, and other information relevant to a customer's decision.
The same pattern appeared among professional services.
Strong agency and architecture websites contained specific services, industries, case studies, project evidence, credentials, awards, and geographic information.
The takeaway isn't “publish more content.”
It's:
Give search engines and AI systems clear, specific evidence about what your business actually does.
Traditional search still matters.
Across several local-service categories, businesses with strong AI visibility also had strong Google Business Profiles, significant review evidence, and competitive Maps visibility.
But we would be careful about interpreting that relationship.
Our research does not establish that ranking higher on Google causes ChatGPT to recommend a business. Web-enabled AI systems can also consult search engines and web results when constructing an answer, making the relationship difficult to isolate.
What the evidence does support is simpler:
Strong Google visibility and strong AI recommendation visibility frequently coexist, particularly for local consumer services.
For businesses, that makes treating SEO and GEO as completely separate strategies increasingly difficult to justify.
This may be the most useful finding of the study.
There doesn't appear to be one universal GEO checklist.
For Portland plumbers and dentists, local reputation was particularly prominent. Reviews, Maps visibility, business profiles, service specificity, and location signals all appeared repeatedly.
For architecture and design, the credibility ecosystem looked different. Documented projects, specialization, awards, professional organizations, press, and industry recognition became much more important.
For personal injury law, reviews existed alongside practice-area pages, legal directories, professional recognition, and other independent evidence.
For SEO and web agencies, third-party lists, case studies, service specificity, and clear Portland relevance appeared repeatedly.
The mechanism changes, but the underlying principle stays remarkably similar:
AI systems need evidence that connects a business with the thing it's being recommended for.
Schema markup is commonly discussed as a GEO tactic, so we specifically looked at whether the research supported that claim.
It didn't.
That doesn't mean schema is useless. Structured data can provide machines with clearer information about an organization, its services, location, people, and other entities.
But we could not consistently inspect raw structured data across every business in the dataset, and the evidence we collected did not establish schema as a differentiating predictor of Portland AI visibility.
So we're not going to claim that it does.
This study found no defensible evidence that schema markup itself predicts whether ChatGPT will recommend a Portland business.
That's different from saying businesses shouldn't use it.
It means businesses shouldn't expect adding JSON-LD to suddenly make ChatGPT recommend them.
Based on the evidence we collected, our current hierarchy looks like this:
Schema may help clarify those relationships, but this study does not establish it as a direct AI recommendation factor.
GEO doesn't appear to be about finding one trick that makes ChatGPT recommend your company.
It's about building a web presence that becomes increasingly difficult for a machine to misunderstand.
Your website says you're a Portland architecture firm specializing in sustainable commercial projects.
Your portfolio proves it.
AIA Oregon confirms who you are.
Industry publications discuss your projects.
Other websites mention your work.
Your business information consistently connects you to Portland.
Search results reinforce those relationships.
Now there isn't one webpage claiming you're relevant.
There's a network of evidence.
That's the central pattern we found across Portland industries:
Businesses appear to become more AI-recommendable when multiple independent parts of the web agree about who they are, where they operate, what they're good at, and why they're credible.
That's also why we don't think GEO should be treated as a replacement for SEO.
Strong websites, useful content, local visibility, reviews, authority, PR, third-party mentions, and clear business information all reinforce each other.
Search is fragmented.
Your visibility strategy shouldn't be.
This was an exploratory Belorado GEO study of six Portland industries conducted in August 2026.
We constructed 60 commercial-intent recommendation prompts across architecture, interior design, personal injury law, dentistry, plumbing, and SEO/web agencies. Recommendation results were used to identify recurring businesses for deeper investigation.
Because the exploratory recommendation test occurred in an environment containing prior Portland-business context, recommendation frequency should be treated as pilot data rather than a clean-room benchmark.
Signal conclusions were based on independently observable evidence surrounding surfaced businesses, including local profiles, reviews, websites, service pages, directories, publications, awards, professional organizations, projects, case studies, and other third-party sources.
The relationships discussed in this article are correlations, not established AI ranking factors. AI recommendations are dynamic, model-dependent, query-dependent, and can change over time.
