94% of business buyers used ai to research their last purchase. your customers are doing it to you right now.
most conversations about AI adoption are about what AI can do inside your business. here's the half nobody's covering: your own customers are already using it to decide whether you make their shortlist, and that decision is happening before they ever contact you.
Forrester surveyed nearly 18,000 global business buyers for its 2026 Buyers' Journey Survey. 94% used AI during their most recent purchase, up from 89% in 2025. More specifically: 55% compared vendors inside AI tools, 54% researched products, and 47% built their internal business case, all before any vendor contact.
And AI answer engines now outrank vendor websites, product experts, and direct sales contact as the most meaningful research source. Twice as many buyers named AI their most meaningful channel over any alternative.
AI vendor research is defined as the practice of using AI assistants to compare suppliers, evaluate products, and build an internal business case before contacting any vendor directly. Forrester's 2026 survey of 18,000 buyers found 94% did some form of it during their most recent purchase.
what this actually changes
For twenty years, a buyer's research path ran through you. They found your site, read your case studies, maybe talked to someone, then checked references. You had multiple chances to establish credibility and correct anything they'd gotten wrong.
That path has moved. Under the AI research model, a buyer's impression of you is largely formed before any contact happens, inside answers you cannot see or directly control. Forrester's own finding is that buyers then validate what AI told them with peers and industry contacts, not with vendors.
Here's the one line worth remembering: the shortlist is now assembled before the first conversation, and if you are absent from it, you never learn you were considered.
The traffic side of this is already visible. B2B companies are reporting website traffic declines of 10 to 40% as research migrates into AI answer engines. That traffic isn't lost to a competitor's website. It's being absorbed by an answer that names three vendors, and the other forty in the category get nothing.
why small operators are more exposed, not less
The instinct is to read this as an enterprise problem. It isn't, and the reasoning is straightforward.
A large, well-known company has years of press coverage, analyst mentions, review-platform presence, and third-party writeups feeding the models a coherent picture of who it is. A small operator frequently has a website, a couple of social profiles, and nothing else. When a model tries to answer "who should I hire for this," it has almost nothing to work with on your behalf.
Harvard Business Review reported in its March-April 2026 issue that LLM data about brands is frequently incomplete or incorrect, and that two-thirds of Gen Z and more than half of Millennials have already started using LLMs to research products. Brands that are miscategorized, absent, or wrongly described have no early-stage defense, because they aren't in the conversation to correct it.
the 5-minute test
You don't have to guess at where you stand. Do this today:
- Open ChatGPT and turn on temporary chat, so your own history doesn't skew the result.
- Ask it to recommend a business like yours, described the way a real customer would describe their problem, not the way you describe your service. Include your area if you serve one.
- See whether you appear.
- Then ask one more question: which sources did it use to pick those businesses?
That last step is the one people skip, and it's the most valuable. The model will list the directories, review platforms, and sites it drew from. That list is your roadmap. It just told you exactly where you need to be and are not.
what actually moves this
Four things, in rough order of how quickly they pay off for a small operator:
- Say exactly what you do and who for, in plain language, in text. "Serving the region" gives a model nothing. "We do X for Y kind of business in these specific places" gives it something to match against a real query.
- Get your details identical everywhere. If your business name, address, and description appear three different ways across the web, a model may not connect them into one trusted entity at all.
- Earn reviews that mention specifics. "Great service" is unusable. A review naming the specific problem you solved is retrievable and quotable.
- Show up somewhere you don't own. This is the biggest lever and the one most often skipped. Analysis of over a million AI prompts by Muck Rack found more than 85% of non-paid AI citations come from earned media rather than vendor websites. One credible mention on a site you don't control outweighs a lot of pages on one you do.
the connection to the rest of your AI work
There's a symmetry here worth naming. The same operators who haven't yet gotten AI working properly inside their business are also, usually, the ones AI can't find when their customers go looking. Both problems have the same root: nobody set up the context the system needs to do the job.
Internally, that means Claude starting every conversation not knowing your business. Externally, it means an answer engine describing your category without you in it. Same gap, two directions.
not sure which gap to close first?
the free instant system finder takes a plain description of what you sell and hands back a paste-ready system to start with, right on the page.
find what to hand ai firstfrequently asked questions
What percentage of B2B buyers use AI for vendor research?
94%, according to Forrester's 2026 Buyers' Journey Survey of nearly 18,000 global business buyers, up from 89% in 2025. Specifically 55% compare vendors in AI tools, 54% research products, and 47% build their internal business case.
Are AI answer engines really more influential than sales reps?
In that survey, yes. AI answer engines outranked vendor websites, product experts, and direct sales contact as the most meaningful research source, with twice as many buyers naming AI over any alternative.
How is this affecting website traffic?
B2B companies reported traffic declines of 10 to 40% as research migrated into AI answer engines. That traffic is not moving to a competitor's website, it is being absorbed by answers that name a small number of vendors.
Why are small businesses more exposed than large ones?
Because a large company usually has years of press, analyst coverage, and third-party writeups giving models a coherent picture of it. A small operator frequently has a website and little else, leaving a model almost nothing to work with.
How do I check whether AI recommends my business?
Open a temporary AI chat, describe your customer's problem the way they would, and see whether you appear. Then ask which sources it used. That second question returns the list of places you need to be and are not.
written by elisabeth hitz, certified in anthropic's ai fluency program (framework & foundations, and ai capabilities & limitations), plus claude 101 and claude cowork. sources: forrester 2026 buyers' journey survey (18,000 global business buyers, published january 2026), via machine relations research; muck rack / generative pulse "what is ai reading?" analysis of over 1 million ai prompts, 2026; harvard business review, "preparing your brand for agentic ai," march 2026. published 26 july 2026.