I spend a lot of time these days looking at the same question from different angles:
Why does one company consistently show up in ChatGPT, Gemini, Google AI Overviews and other AI-generated answers, while another seemingly similar company does not?
The answer is rarely one tactic.
It isn’t because somebody added an FAQ. It isn’t because they installed a new SEO plugin. It isn’t because they created an llms.txt file or changed a few title tags.
And it definitely isn’t because they figured out the magic prompt that makes ChatGPT love them.
Recently, Overit CMO Sean Fitpatrick and I created and presented a series of webinars about gaining visibility in LLNMS, also known as the process of “AISEO.” As we have analyzed visibility across industries including Home Services, Financial Services and Healthcare, some pretty clear patterns have emerged. The details change by industry, but the underlying signals are remarkably similar.
That is encouraging because much of what works is still rooted in good, ol’ SEO—just applied more strategically to the way people search and the way AI systems assemble answers today. Our AISEO framework has centered around five areas: authority topics, clear and helpful content, conversational language, stronger schema and being present in the places AI systems seek information.
But there are a few additional things we are learning along the way.
1. High AI Visibility Starts With Real Topical Authority
One of the biggest mistakes I see is treating AI search like traditional keyword SEO with a new name.
Find a prompt. Optimize a page for the prompt. Track whether you appear.
That is probably too narrow.
Recent research from Semrush involving 50,000 brands reinforces something we are seeing in practice: AI visibility is increasingly a topic-level competition. Winning one prompt is not the same as being consistently associated with the topic.
This changes how I think about keyword research.
I still want search volume. I still want rankings. But I am increasingly interested in the collection of questions someone asks while making an actual decision.
For a home services company, that might include:
- What is causing this problem?
- Is it dangerous?
- Can I fix it myself?
- How quickly do I need somebody?
- How much might the repair cost?
- What kind of professional should I call?
- Who provides that service near me?
A financial services customer may move through questions about account eligibility, contribution limits, investment choices, taxes, custodians, fees and eventually provider comparisons.
A healthcare journey can move from symptoms to possible conditions, testing, treatment options, specialists, locations and what to expect during recovery.
Those are not isolated keywords. They are topic ecosystems.
This is why we have sometimes recommended consolidating overlapping educational content rather than continuing to produce another page for every slightly different keyword variation. Fewer, stronger content ecosystems can create a much clearer signal about the subjects a brand actually deserves to be considered authoritative about.
The question I keep coming back to is the same one from our AISEO presentations:
Where do you really deserve to be at the top of search?
Start there.
2. The Best Content Is Easy for Humans—and Machines—to Extract
There was a period in SEO when longer often seemed better.
If the competition had 1,500 words, maybe we needed 2,000.
Thankfully, that was never a particularly great content strategy, and it becomes even less useful when optimizing for AI answers.
High-visibility content tends to make its information extremely easy to identify.
That means:
- Clear headings
- Shorter sections
- Direct answers
- Helpful summaries
- Lists where lists make sense
- Tables where comparisons make sense
- Definitions that actually define things
- Important facts that don’t require reading six introductory paragraphs first
Our webinar guidance describes this as making content helpful, summarized, bite-sized, skimmable and conversational.
There is now outside data behind this as well. Semrush’s analysis of AI citations found positive associations with clarity and summarization, E-E-A-T signals, Q&A formatting, section structure and structured data elements.
That does not mean every page should turn into a giant FAQ.
It means the page should make individual ideas easy to understand and retrieve.
We see opportunities for this constantly.
A home services page explaining a seasonal problem should quickly explain why it happens, what homeowners will notice and what they should do next.
A retirement account page should not bury eligibility requirements or contribution information beneath 800 words of general retirement planning copy.
A healthcare condition page should make it easy to understand symptoms, risk factors, diagnosis and treatment without forcing patients to translate medical marketing language.
The best AISEO content is usually better human content first.
3. Conversational Search Means We Need to Answer Better Questions
People talk to AI differently than they type into Google.
Someone might Google:
chimney repair albany
But ask an LLM:
My chimney has cracks near the top but it isn’t leaking yet. Is this something I need repaired before winter?
The underlying need may be similar. The language and context are very different.
Financial queries become even more interesting:
Can I invest retirement money in real estate without personally taking possession of the property?
Healthcare might look like:
What type of doctor should I see if I’m getting shortness of breath and irregular heartbeats during exercise?
These prompts contain context, qualifiers and implied follow-up questions.
That makes traditional keyword research only part of the process.
We have been combining keyword data with People Also Ask questions, related searches, LLM prompt research, customer questions, sales-team input and subject-matter expertise to understand the language around the decision, not just the keywords associated with it.
FAQs can help, but they should be genuine questions relevant to the specific page—not a generic block pasted across an entire website. And useful conversational answers do not always need to literally appear in question-and-answer format.
Write like somebody asked you the question.
Then answer it.
4. Entity Clarity May Be One of the Most Underrated AISEO Opportunities
This is where schema becomes much more interesting than simply checking whether a website “has schema.”
Most sites do.
A typical WordPress plugin can provide WebSite, Organization, WebPage, BreadcrumbList and some basic author information. That is a good foundation.
But it doesn’t necessarily explain the business particularly well.
We have been building more advanced schema graphs that connect the actual entities and concepts represented on a page.
In Financial Services, that can include types such as FinancialService, InvestmentOrDeposit, LoanOrCredit, FAQPage, HowTo, Dataset and CollectionPage. In Healthcare, the vocabulary can become even more descriptive through entities such as MedicalCondition, MedicalProcedure, Physician and MedicalOrganization.
Home Services has its own opportunities around the business, locations, services, service areas, reviews and the relationships between them.
The point isn’t to see how many schema types we can stuff onto a page.
It is to clearly tell machines:
This is who we are. This is what this page is about. These are the services, products, people, places and subjects connected to it. And this is how those things relate to one another.
That becomes particularly useful for companies with multiple locations, multiple brands or complicated service offerings.
Schema helps AI systems understand content; it does not magically turn weak content into authority.
Think of it as reinforcing the story the website is already telling.
5. Your Website Is Only One Source in the Answer
This may be the biggest mental shift.
For years, SEO teams understandably focused primarily on improving the website.
AI systems are forcing marketers to zoom out.
They can build an understanding of a company from information found across the web: media coverage, associations, partner sites, reviews, social networks, video platforms, professional profiles, forums and other third-party sources.
Ahrefs found branded web mentions to be strongly correlated with brand appearances in its large-scale AI Overview research.
This has affected several strategies we have worked on.
For Home Services businesses, we have looked beyond traditional link building toward distributors, manufacturers, associations, local organizations, charities and other entities that can independently validate what the business does and where it operates.
For Financial Services companies, we have examined how related corporate websites, partners and authoritative educational sources can reinforce the connection between a brand and very specific financial topics.
In Healthcare, credible outside references become particularly valuable. Provider affiliations, health systems, universities, government resources, professional organizations, media appearances and expert commentary all help reinforce expertise in ways a self-written service page cannot do by itself.
Even LinkedIn deserves a different look. Semrush analyzed 89,000 LinkedIn URLs cited by AI search platforms and found LinkedIn appearing in about 11% of AI responses in its dataset.
The broader lesson is simple:
You cannot establish all of your own authority.
Other people and organizations have to help establish it for you.
That is why PR, social, video, partnerships, digital marketing and SEO are becoming increasingly difficult to separate.
6. Consistency Across Entities Matters More Than Ever
Another pattern that becomes especially obvious in multi-location businesses is inconsistency.
The website calls the company one thing.
Google Business Profile calls the location something slightly different.
A directory uses an old brand name.
The physician profile uses one specialty term while the hospital website uses another.
An acquired company still has dozens of strong third-party references under the old brand.
Two related corporate websites describe the same financial offering differently.
Traditional SEO has dealt with versions of these problems forever.
With AI systems attempting to reconcile information across many sources, I believe entity consistency becomes even more important.
This doesn’t mean every mention across the Internet has to be identical.
It means the relationships should make sense.
For multi-location Home Services brands, we have spent time aligning location naming, service-area pages, Google Business Profiles and structured data.
In Healthcare, the same concept applies to physicians, specialties, facilities, conditions and treatments.
In Financial Services, it can involve brands, account types, investment categories, educational resources and even multiple domains owned by the same organization.
The easier it is for an AI system to determine who you are and what you are authoritative about, the better starting point you have.
7. A Citation Is Not the Same Thing as Brand Visibility
This is an important distinction as AISEO reporting becomes more sophisticated.
Getting your URL cited feels great.
But did the AI actually mention your company?
Not necessarily.
A 2026 Semrush study found that 61.7% of AI citations were “ghost citations”—the source was cited, but the brand itself was not named in the answer. The same research found comparative content generated substantially more brand mentions than purely informational content.
That suggests we need to start separating at least two goals:
Citation visibility: Is our information being used?
Brand visibility: Is the company actually becoming part of the answer or recommendation?
Both matter.
But they are not the same KPI.
For brands that have already built strong educational libraries, this can be an especially interesting opportunity. The content may already be influencing AI answers even when the brand is receiving very little credit for it.
That is a different problem than simply “we need more content.”
8. Stop Thinking About “Ranking in ChatGPT”
I hear variations of this all the time:
How do we rank #1 in ChatGPT?
We probably need to retire the question.
There isn’t one ChatGPT result, just as there isn’t one universal AI search result.
Prompts change answers. Geography can change answers. Personalization can change answers. The model can change. Retrieval sources change.
Even different reasoning modes within ChatGPT can produce dramatically different source sets. Semrush recently found only 25.6% overlap between cited domains in minimal- and high-reasoning modes for the same prompts.
And AI citations themselves can be volatile. One longitudinal study following 1,127 cited URLs found that 66% of URLs cited during its first collection period were no longer cited four weeks later.
That doesn’t make optimization pointless.
It changes what success should look like.
Instead of asking whether we rank for one prompt today, I would rather know:
- How frequently are we mentioned across an important topic?
- How often are competitors appearing instead?
- Which pages and outside sources are influencing those answers?
- Are we being cited, mentioned or recommended?
- Which entities does AI associate with our brand?
- Where are the gaps in that understanding?
- Is visibility improving across a meaningful collection of prompts over time?
That is a much more useful AISEO program.
Different Industries. Surprisingly Similar Fundamentals.
Home Services, Financial Services and Healthcare could hardly be more different.
The regulations are different. The customer journeys are different. The level of risk is different. The content is different.
But when we look at websites and brands earning strong visibility in AI answers, the fundamentals keep converging.
They tend to:
- Own specific topics instead of chasing everything.
- Answer real questions clearly.
- Structure information so both humans and machines can understand it.
- Define their entities and relationships accurately.
- Earn validation outside their own websites.
- Maintain consistency across their digital presence.
- Measure visibility across topics instead of obsessing over a single prompt.
That sounds a lot like advanced SEO, content strategy, digital PR, brand marketing and technical optimization working together.
Which is probably the most important cross-industry trend of all.
AISEO isn’t replacing SEO.
It is forcing us to finally stop treating SEO like an isolated channel.
