The shortlist moved, and nobody told you
For years the buyer’s journey started with a search and a few open tabs. Now it starts with a question typed into an assistant: “who are the best firms for this, and why.” The machine answers in a paragraph, names a handful of companies, and the buyer treats that paragraph as a starting line. Forrester’s survey of 18,000 buyers found 94% used AI in their most recent purchase, and more than half compared vendors inside those tools before speaking to a single sales team. By the time an enquiry lands, the decisive round is already over.
Here is the part that should worry the quiet companies. In the same research, roughly a third of buyers ended up choosing a vendor they had never heard of before the machine introduced them. The shortlist is no longer a reward for being established. It is a live question, re-answered every time someone asks.
Why buying your way in won’t work
The industry’s response has been predictable and fast: a whole market of “GEO” tools promising to get you cited in AI answers, the way SEO once promised the top of the page. Some of the plumbing is real. But treat GEO as a hack and you will be disappointed, because an AI recommendation is not a slot you can win. It is a summary of what the rest of the world already says about you.
These systems don’t invent opinions. They compress them. They read the reviews, the mentions, the comparisons, the podcasts, the posts, and hand back the consensus in a confident voice. If the consensus about your category doesn’t include you, no amount of tagging fixes that. You can’t optimise your way into a conversation you were never part of.
What the machine actually rewards
Strip away the novelty and the answer is old. The machine recommends the same companies a well-connected human would: the ones that are easy to recall and easy to vouch for. Marketers have a plain name for the first half of that, mental availability — the likelihood you come to mind in a buying moment. It is built by showing up consistently, in a recognisable way, long before anyone is ready to buy. What was true for human memory turns out to be true for the machine’s.
The second half is proof. A machine, like a cautious buyer, wants corroboration before it puts its name to a recommendation. Case studies with numbers. Reviews that read like real people. The brands that dominate AI answers aren’t the ones shouting loudest. They’re the ones the rest of the web keeps agreeing about.
Your social channel just got a second job
This is where “we post now and then” quietly costs you. Social media used to have one visible job: reach the humans scrolling. It now has a second, invisible one. Every post, every comment thread, every time a client mentions you, is a small deposit into the evidence these systems read to decide who’s credible. A distinctive, proof-led feed doesn’t just win the person watching. It teaches the machine what to say when someone asks.
Where to start
Open ChatGPT and ask it, as your buyer would, who the best options in your category are. Read the answer coldly. If you’re missing, you’re not being penalised by an algorithm — you’re being accurately described as unknown. That gap is fixable, but not with a plug-in. It’s fixed by a social system with a clear position, distinctive assets, and proof in every post, run long enough to be believed. That’s the whole point of the Social Strategy Sprint: to make you the name that comes up, whether the one asking is a person or the machine they now ask first.