Last week I stood up in a meeting and said something I believe pretty strongly:
AI visibility right now feels a lot like SEO did in 2007 or 2008.
Back then, the field was still open.
Most business owners knew Google mattered, but relatively few understood what made one website easier for Google to crawl, understand, and rank than another.
That created an unusual opportunity.
You did not always need the biggest company, the biggest marketing budget, or the prettiest website.
Sometimes you could beat a much larger competitor simply because you got the fundamentals right before they did.
- You made the site easy to crawl.
- You organized the pages correctly.
- You clearly explained what the business did.
- You created useful content.
- You made locations and services obvious.
- You earned links and references from other sites.
None of those things were magic.
The advantage came from doing the work while most competitors were barely thinking about it.
Eventually everybody caught on. SEO became an industry. Competitors improved. Search engines became much more sophisticated. Getting the basics right was still important, but the easy advantage disappeared.
I think we may be standing in a similar window again.
Your website has another audience now
People still make buying decisions. That has not changed.
But more systems are helping people decide which companies they should even consider.
Google Search already uses AI-driven systems to retrieve and synthesize information, including techniques such as query fan-out and retrieval-augmented generation. Google still says the fundamentals of SEO remain important for these experiences.
Then there are tools like ChatGPT, Perplexity, AI search systems, recommendation engines, and a growing category of browser agents that can research and interact with websites.
That changes the job of the website.
A website still has to earn human trust. It also increasingly has to give machines enough clear information to understand the business.
That means an intelligent system may need to figure out:
- What does this company actually do?
- Who does it serve?
- Where does it operate?
- Which specific services or products does it offer?
- What makes the company credible?
- What evidence supports those claims?
- Is the information current?
- Do other sources agree?
- How should someone contact or do business with the company?
A website can look excellent and still answer those questions poorly.
That is the opportunity.
Schema is useful. It is not the whole answer.
One of the shortcuts I expect us to hear more often is:
"We added schema, so the website is ready for AI."
Schema matters. We use it. Structured data can help machines understand facts and relationships more explicitly.
But schema is one layer.
A business can have technically correct schema while its service pages remain vague, its credentials are buried, its locations are unclear, its reviews are disconnected from the claims it makes, and outside sources tell a slightly different story.
Our own approach at Two Tall treats AI visibility as a broader trust system. Schema sits alongside clear content structure, entity clarity, reviews, current information, third-party references, topical depth, brand mentions, and external corroboration.
The objective is bigger than marking up a page correctly. The objective is to make the real company easier to understand and verify.
The technical layer is changing too
This is where things get especially interesting.
Chrome has introduced an experimental Agentic Browsing category in Lighthouse for evaluating how websites behave when software agents interact with them.
Among the things Chrome currently examines are the accessibility tree, page stability, WebMCP-related interaction capabilities, and the presence of an llms.txt resource. Chrome describes llms.txt as an emerging convention for providing a machine-readable summary of a site's content.
That does not mean llms.txt is a Google ranking factor.
Google Search specifically says it does not require llms.txt, Markdown files, special AI markup, or special AI schema to appear in AI Overviews or AI Mode.
Both things can be true. Google Search is one system. Browser agents are another emerging use of the web. And llms.txt itself remains an evolving convention rather than a universal requirement. Its published specification is designed to give language models a concise Markdown overview of a site and selected resources.
Some of today's standards will survive. Some will change. Some may disappear.
That is what an emerging field looks like.
The opportunity is not predicting every winning standard perfectly. The opportunity is building a foundation that makes the business easier for intelligent systems to work with as the standards mature.
Most established businesses already have the raw material
This is the part I think gets overlooked.
For many good businesses, we do not have to manufacture credibility. It already exists.
They already have:
- years of experience
- satisfied customers
- good reviews
- certifications
- specialized expertise
- employees with deep knowledge
- strong products
- valuable partnerships
- completed projects
- industry experience
- service areas they understand extremely well
The issue is often that this evidence is scattered, buried, outdated, vaguely described, or missing from the places machines and buyers use to understand the company.
That creates an asymmetric opportunity.
You may not need to become a better company than your competitor. You may need to become a much easier company to understand.
That is what reminds me of 2007
Early SEO rewarded businesses that understood the new environment before everyone else did.
I believe something similar may be happening now.
The field will get more crowded. Agencies will package AI visibility. Software platforms will automate pieces of it. Businesses will begin competing intentionally for inclusion in AI-assisted answers and recommendations. The technical standards will mature. The inexpensive advantage will get harder.
What we do not know is exactly how long the current window lasts.
What we can see is that the window exists.
A large number of businesses still have websites built primarily as online brochures. Their competitors are starting to be evaluated by systems that want structured, explicit, current, credible information.
That gap creates opportunity.
Five questions I would ask about your business today
You do not need to chase every AI trend. Start here:
Can an intelligent system clearly identify what our company does?
Can it determine exactly who we serve and where?
Is there visible evidence supporting the claims we make?
Do our website, Google presence, reviews, directories, and other public sources agree?
Is the website technically structured so search engines and emerging agents can understand and interact with it reliably?
If those answers are unclear, there is work worth doing.
At Two Tall, this is becoming part of how we evaluate websites. We look at the same fundamental question we have asked for years:
Does the online version of the company accurately represent the business people would encounter in real life?
Now we are also asking:
Can the systems helping those people decide clearly understand that business too?
I think businesses that answer both questions early may have an advantage — just like the businesses that understood SEO before everybody else did.