Whether GEO is worth starting on now or whether the whole thing is a few years premature is a fair question, and I do not think it can be answered by pointing at how many people use ChatGPT. The useful version of the question is what the last shift looked like from the inside, before anyone knew how it turned out. In October 1997, the year search engine optimization got its name, 22.2 percent of Americans used the internet, there was no evidence that a search ranking was worth money to anybody, and the entire field was a handful of practitioners arguing on mailing lists. Anyone who decided search was not worth the effort that year was making a reasonable call with the information available, and that is close to where most people sit on GEO today.
Where the web actually was when SEO got its name
That 22.2 percent comes from Table 2-2 of the Commerce Department's report on Census Bureau survey data, which tracked internet use from any location for everyone aged three and older (NTIA, A Nation Online, 2002). Google had not launched yet, AltaVista and Yahoo were splitting the traffic between them, and a marketing director who wanted to spend money on search rankings had nothing to point at when finance asked what the return would be.
The ramp after that was steady rather than explosive. Internet use reached 32.7 percent by December 1998, 44.4 percent by August 2000, and 53.9 percent by September 2001, so crossing half the country took just under four years from that starting point. Google introduced PageRank in 1998, the first Search Engine Strategies conference ran in 1999, and the tooling slowly assembled itself around an audience that was already there.
Is AI actually moving faster than the web did?
Pew Research put AI chatbot use among US adults at 49 percent in February 2026, up from 33 percent in 2024 and 18 percent in 2023, with about a quarter of adults now using one every day (Pew Research Center, Americans and AI 2026). ChatGPT launched in November 2022, so reaching roughly half of American adults took about three years.
I expected a bigger difference than that, and I believe most of the "fastest adoption in history" writing you see on this topic is comparing against a version of the 1990s that never existed. The web climbed from 22.2 percent to 53.9 percent in just under four years while AI climbed from 18 percent to 49 percent in about three, so AI is covering roughly the same ground about a year quicker, which is faster without being the different-universe leap the hype suggests.
Chart data
| Series | Point | Share |
|---|---|---|
| Americans using the internet | October 1997 | 22.2% |
| Americans using the internet | December 1998 | 32.7% |
| Americans using the internet | August 2000 | 44.4% |
| Americans using the internet | September 2001 | 53.9% |
| US adults using chatbots | 2023 | 18% |
| US adults using chatbots | 2024 | 33% |
| US adults using chatbots | February 2026 | 49% |
The Census surveys counted everyone aged three and older while Pew surveys adults, and the same NTIA report shows how much the answer moves depending on what you measure, since September 2001 reads as 50.5 percent of households, 43.6 percent of people using the internet at home, or 53.9 percent using it anywhere. I picked the measure that matches how Pew counts, and the ramp shapes hold up better than any single number does.
Why did SEO budgets take so long to arrive?
More than half the country was online by September 2001, and SEO still did not become a normal line item on a marketing budget until the middle-to-late 2000s, which is a gap of roughly five to eight years between the audience arriving and the money following. Market-size estimates for the industry are vendor-published and inconsistent enough that I would not stack them into a trend line.
The lag existed because the tooling did not exist yet, since ranking checkers were crude, attribution barely worked, and a marketing director who wanted budget approval had no way to show what organic search was worth, so the spending waited for the measurement to catch up. My issue with most "AI is moving faster" arguments is that they focus entirely on user adoption and skip this second curve, when the second curve is where the opportunity actually sits.
This time the order is reversed. Chatbot use reached roughly half of US adults in early 2026 and the measurement tools were already there, which means the five-to-eight year wait for someone to build the instruments has not happened.
What being early actually buys you
Being early does not buy you traffic, and I would rather say that plainly than sell you a number that will not show up in your analytics. What it buys you is the chance to be described correctly while the models are still forming their picture of you, which is worth more to a young brand than it is to one that has been written about for a decade.
Most companies are being read by crawlers already, without anyone having decided what those pages should say to a machine, and that is a cheaper problem to fix now than after the models have settled on a version of you. If you are still deciding whether the spend makes sense at all, I worked through the cases where the answer is no in whether AI brand visibility is worth your money, and the practical differences between the two disciplines are laid out in GEO, the new SEO.
The part we work on at Ooky is the reading itself. Brand Intelligence serves the crawlers a structured version of your brand at the edge, so GPTBot finds facts you wrote rather than a page it has to guess around. For the fields you cannot afford to have invented, a do_not_infer directive tells the model to leave them blank instead of filling them in.
Brand Tracker then runs the buying prompts on a schedule so you can watch whether the answers shift over time, which is a correlation you can observe rather than a proof of cause.
I do not think GEO is urgent for everyone, and a company with no inbound motion and a long offline sales cycle can reasonably wait. The crawlers are not waiting though, and the numbers above are what they did to a company that had not got round to the question yet.
FAQ
Is it too early to invest in GEO?
It is early in the sense that budgets and job titles have not settled yet, and it is not early in the sense that the crawlers are already reading your site. Pew put US adult chatbot use at 49 percent in February 2026, up from 18 percent in 2023, so the audience has already arrived even though the discipline around it has not.
How fast did AI adoption move compared to the early web?
The two curves are closer than most people assume. Internet use among Americans went from 22.2 percent in October 1997 to 53.9 percent in September 2001, which is just under four years, while AI chatbot use among US adults went from 18 percent in 2023 to 49 percent in February 2026, which is about three years.
What is the difference between GEO and SEO?
SEO works to rank a link on a results page, while GEO covers whether your brand gets cited, recommended, and described accurately inside a generated answer. Accuracy is only one part of it, since being left out of the shortlist costs you just as much. The two overlap, because a model that cannot crawl you cannot cite you either, but the measurement and the fixes are not the same.
See what the crawlers are already taking from you
Ooky logs every AI bot that visits your site, serves them a structured version of your brand, and tracks whether ChatGPT, Claude, Gemini, and Perplexity start describing you accurately. The free tier covers the tracking, so you can see the gap before you spend anything.
Sources
- National Telecommunications and Information Administration, US Department of Commerce. "A Nation Online: How Americans Are Expanding Their Use of the Internet." February 2002. Web adoption figures are from Table 2-2, "Internet Use From Any Location by Individuals Age 3 and Older," covering October 1997, December 1998, August 2000 and September 2001, based on Census Bureau Current Population Survey supplements. Retrieved 2026-07-29. https://www.ntia.gov/sites/default/files/ntia/publications/anationonline2.pdf
- Pew Research Center. "Americans' Views on AI Chatbots, Smart Devices and AI's Impact." June 17, 2026. Survey fielded February 17-23, 2026, n=5,119 US adults, margin of error ±1.6 percentage points. Retrieved 2026-07-29. https://www.pewresearch.org/internet/2026/06/17/americans-and-ai-2026-chatbots-smart-devices-and-views-on-impact/
- Pew Research Center. "Internet use" (long-term trend chart). Note that Pew began systematically tracking US internet use in early 2000, by which point about half of adults were already online, which is why the pre-2000 figures in this post come from Census Bureau data instead. Retrieved 2026-07-29. https://www.pewresearch.org/chart/internet-use/