Somewhere in Michigan this morning, a facilities manager needed a new HVAC contractor. She didn’t Google it. She opened ChatGPT and typed: “Who are the best commercial HVAC contractors near Traverse City for a 40,000-square-foot building?” Thirty seconds later she had three names, a note about response times, and a suggestion on what to ask about maintenance contracts. She called one of the three.
She never saw a search results page. She never clicked an ad. And the contractors who weren’t in AI’s answer? Their websites, their reviews, their thirty years of reputation – none of it was in the room when the decision was made.
This is not a prediction about where things are headed. It’s a description of any given day. And it’s the most important thing happening to small business right now because it sits upstream of everything else. We spend a lot of energy talking about how AI can streamline our operations – and it can – but efficiency is meaningless if your buyers can no longer find you. A business that automates its back office while disappearing from the conversations where customers choose is just a very efficient ghost.
Let me show you how fast this is moving, and where it goes next.
The Fastest Behavior Change in the History of Technology
For twenty years, digital marketing had a single tent pole: Google. You made content, Google sent you clicks, you converted the clicks. Every marketing dollar you’ve spent since 2005 – the website, the SEO, the ads – assumed that arrangement was permanent.
It’s being dismantled in real time.
ChatGPT reached roughly 900 million weekly users this year and became the fastest application in history to hit a billion monthly users – faster than TikTok, faster than Instagram, faster than Google Maps. Google’s own assistant, Gemini, is approaching a billion users itself. This is not one upstart eating Google’s lunch; it’s discovery fragmenting across a handful of AI assistants, each of which surfaces, ranks, and recommends businesses differently. The single algorithm every marketer optimized for has become five or six, and none of them work like the old one.
Business buyers moved first and fastest. Buyers now rank AI assistants among their most meaningful sources for researching vendors – ahead of vendor websites and sales reps. The comparing, the shortlisting, the “Who should I even call?” increasingly happens inside an AI conversation before anyone at your company knows the buyer exists.
Here’s the number that should reframe your marketing budget: when Google shows an AI-generated answer at the top of the page, only about 8% of people click through to any website at all, according to Pew Research. Google isn’t dying; it still handles billions of queries a day. But its role is changing from a router that sent you traffic to an answer engine that keeps users on its own page. The click, the thing we all spent two decades optimizing for, has decoupled from the decision. Your traffic chart can look stable while the decisions quietly migrate somewhere your analytics can’t see.
Machines Are Now Your Majority Visitor
This year, for the first time in the history of the internet, machines generate more web traffic than humans. Cloudflare, which handles a meaningful share of all internet traffic, reported that bots – AI assistants reading pages, comparing suppliers, and gathering answers on people’s behalf – now outnumber the rest of us.
That traffic isn’t junk. Adobe found AI-driven traffic to U.S. retailers grew 393% in a single quarter, and the humans who arrive at a website after an AI conversation convert better than traditional visitors. They show up pre-researched, with a shortlist of one. The research chores that used to generate your website traffic, like twenty open browser tabs or comparison spreadsheets, are being outsourced to machines. Nobody loved doing that work. It won’t argue on its way out; it will just leave the same way users left the phone book.
Now the uncomfortable part. Adobe also ran thousands of web pages through AI readers and found that roughly a quarter of homepage content, and a third of product-page content, can’t be read by AI at all. The facts are locked inside images, buried in design, or written as slogans. “Quality you can trust since 1987” means something to a person standing in your lobby, but to an AI deciding which three contractors to recommend, it means nothing. It needs facts it can find and verify: what exactly you sell, real specs, real pricing, service areas, lead times, and proof.
And you will never see the failure happen. There is no error message when an AI reads your site, finds nothing usable, and recommends your competitor. No bounce appears in your analytics. You just quietly stop appearing in the conversations where buyers make decisions.
Next: Machines That Don’t Just Recommend – They Act
Everything above describes AI as an advisor: it informs a human, and the human acts. The next phase is already emerging, and it’s bigger. AI agents don’t just answer questions, they do things. They compare suppliers, request quotes, check availability, book appointments, and place orders on behalf of their human masters. McKinsey, a strategy and management consulting firm, projects agentic commerce – purchases researched and executed by AI on a person’s behalf – will grow into a trillion-dollar channel within a few years.
Think about what that means for a distributor, a contractor, or a professional practice. Today the question is: “When a machine reads my website, what does it learn?” The question arriving right behind it is: “When a machine tries to buy from me, can it?” Can an agent get a quote from you without a phone call? Check whether you serve its owner’s zip code? Confirm a spec, a price, a lead time? If the answer is no – and today, for most small businesses, it’s a no – the agent moves to the competitor where the answer is yes. Machine customers don’t have brand loyalty, and they don’t leave voicemails.
Every prior digital shift gave us years to adjust. The web took a decade to become mandatory. Social took most of another. This one is compressing the same transition into quarters, because AI adoption is compounding on infrastructure (smartphones, cloud, connectivity) that’s already everywhere.
What to Do This Quarter
None of this requires becoming a technology company. It requires seeing clearly and starting. Four moves to take action:
Ask an AI about your business today. Open ChatGPT or Claude and ask the question your best customer would ask. Are you in the answer? Is what it says about you accurate? That’s your new storefront, and most owners have never looked at it.
Audit what a machine can read. Look at your website the way an AI does; not “How does it look?” but “What does it learn?” Are your services, specs, pricing, service areas, and proof written as verifiable facts or locked in images and slogans? Your decades of expertise sitting in PDFs and brochures is invisible to the systems now doing the choosing. Getting it out is the highest leverage marketing work available to you right now.
Make it possible for machines to act. Inventory every step a buyer must complete by phone or email, such as quotes, availability, or scheduling. Each one is friction a human tolerates and an AI agent won’t. You don’t need to fix them all this quarter. You need to know where they are and start.
Change the question you ask your marketing partners. Stop asking about rankings and impressions. Ask: “Where do we appear when buyers ask AI assistants about our category, and what’s the plan to improve it?” A vague answer is itself an answer.
One closing thought: Your buyers didn’t hold a meeting and decide to abandon the way they found you for twenty years. They just started asking better questions of better machines, one search at a time, like when shoppers drifted from Main Street to the web while the signage out front stayed exactly where it was. The traffic doesn’t announce it’s leaving. The owners who come out ahead of this shift won’t be the ones who waited for certainty. They’ll be the ones who looked at their new storefront – the AI answer – before their competitors knew it existed.
By Alex Belding, Co-Founder of WebriQ; originally published in SBAM’s September/October 2026 issue of FOCUS magazine.
Click here for more News & Resources.