Here’s the short version:
AI is now recommending brands. Here’s why inconsistent publishers improve your ai brand visibility.
Picture the buying process for your last big supplier decision. Somebody on the team opened ChatGPT and typed something like “best video production companies for B2B in the North East” or “content agencies that work with financial services firms”. They got back a paragraph and four or five names. Maybe yours. Probably not.
That’s the shift. For twenty years the game was ranking on a results page a human would scroll through. Now a growing share of buyers never see the page. They see a synthesised answer, and the AI has already done the shortlisting for them.
Which raises the only question that matters: what makes an AI engine trust a brand enough to recommend it?
Where do AI engines get their information about your brand?
Not from your homepage, or at least not mainly. AI answer engines read the whole web and weigh sources against each other. Your own site is one input among many, sitting alongside YouTube, LinkedIn, Reddit, review platforms, press coverage and anywhere else your name shows up.
The most useful public data on this comes from VaynerMedia’s 2026 research with Profound, an AI visibility platform. Across the six major answer engines they studied, social, creator and user-generated content grew as a share of citations on five of them over six months. Different engines lean on different platforms. Their findings showed Google’s AI products drawing heavily on YouTube, ChatGPT leaning on Reddit, and Copilot pulling from professional networks like LinkedIn.
Two things follow from that. There’s no single platform to win, so anyone selling you a one-trick “AI SEO” fix is selling snake oil. And third-party voices matter more than your own claims. A client explaining on camera why they hired you carries more weight with these systems than your services page saying you’re brilliant. That’s uncomfortable if your marketing has always been you talking about yourself. It’s also exactly how buyers have always made decisions, the machines have just caught up.
Does AI prefer video?
The strongest single signal in the Vayner and Profound data was YouTube’s climb. As a citation source it moved from roughly 247th to 34th in ChatGPT, and from 61st to 13th in Copilot. No other platform moved like that.
Why video? Because answer engines favour content that explains, compares and teaches over content that promotes. Reviews, tutorials, walkthroughs, honest comparisons. A two-minute video of your operations director explaining how you handle a compliance question does more for your AI visibility than a month of “we’re delighted to announce” posts, because it’s the kind of source an engine can quote when a buyer asks how that problem gets solved.
For B2B brands this is quietly good news. You don’t need to dance on TikTok. You need people who know their subject, on camera, explaining things clearly, published somewhere AI reads. YouTube sits at the top of that list, and most B2B firms in this region still treat it as an afterthought or a dumping ground.

How often do you need to publish?
Here’s the number that should change how you plan. In the same study, the median time from publishing a piece of content to it being cited by an AI engine was 6.8 days, and 90% of cited content got picked up within about 37 days. But half of everything cited was under 13 weeks old.
Sit with that. The window of content AI draws on slides forward every single day. Whatever’s working for you right now will age out of it within a quarter. A brand that publishes in bursts, a campaign in spring, silence until autumn, keeps falling out of the answer and starting again from nothing.
Most marketing teams still run on campaign timelines. AI engines don’t. They reward the brand that shows up every week over the brand that shows up twice a year with something polished. Consistency has stopped being a nice-to-have for brand building and become a technical requirement of how these systems retrieve information.
We’d say that, obviously. B7’s whole model is built on quarterly production feeding a steady publishing rhythm. But we built it that way because sporadic content never worked for buyers either. The machines are just enforcing a rule humans already lived by.
What should a B2B brand actually do about this?
Skip the exotic tactics for now. The basics carry most of the weight, and most firms haven’t done them.
Pick a cadence you can sustain and hold it. Weekly beats a monthly burst, and a monthly rhythm you keep beats a weekly one you abandon in March. The 13-week window punishes gaps more than it rewards volume.
Put your experts on camera answering real questions. The questions clients ask you on calls are the questions buyers type into ChatGPT. Answer them plainly, in video and in writing, with the person who actually knows.
Treat YouTube as a search channel. Proper titles that match how people phrase questions, real descriptions, and transcripts. It’s the fastest-rising source in the data and the least crowded space in B2B.
Keep your site open and boring in the right ways. Visible publish dates, named authors, services and pricing information AI can actually read, nothing important gated behind a form. An engine can’t cite what it can’t see.
Earn mentions you don’t control. Client case studies in the client’s words, guest appearances, reviews, industry press. The engines trust others talking about you more than you talking about you.
The honest bit
Nobody outside these companies fully knows how the weighting works, and it’ll change. Anyone promising you a guaranteed spot in ChatGPT’s answers is guessing with confidence.
What the data does show is that the behaviours these systems reward were already the behaviours of good marketing: publish regularly, explain rather than promote, let real people and real clients do the talking. If you’ve been putting that discipline off, AI recommendations are the deadline it just acquired. The brands teaching these systems who to trust are doing it now, this quarter, while the shortlists are still being formed.
If you want to know whether AI engines mention your brand today, the test costs nothing. Open ChatGPT and ask it the question your best prospect would ask. If you’re not in the answer, you’ve found this quarter’s job.
Find out how B7 can help you with your AI Brand Visibility →
