A prospective customer asked an AI about you. It gave a confident answer. Some of it was right. Some of it wasn't. You weren't in the room — but the decision was already being shaped.
Purchasing decisions are increasingly moving upstream. Whether personal or B2B, we're all using AI for discovery and research. And the reality is we love it — it saves time, surfaces things we wouldn't have found otherwise, and often knows more about a product than the product's own site appears to. That's not changing.
For businesses though, AI-referred traffic is a black box. Roughly 7% of AI sessions result in an outbound click, which means most buyers who encounter you in an AI conversation will close the chat and — maybe a day later, maybe a week — find their way to you through some other channel. The good news is that the traffic we can track from AI converts at multiples higher than organic search. That makes sense: by the time they reach you, the AI has already done the selling.
But the attribution challenge extends further than missing click-throughs. What happens after the chat closes is where it gets complicated. The buyer learns about your brand in AI, closes the window, checks your social, Googles your name because they don't know your URL, maybe asks the AI again the next day. A myriad of paths. The channels that eventually get credit — SEO, paid, social, direct — are increasingly downstream of where the actual decision formed. That attribution picture is becoming less and less accurate, and most companies don't know it yet.
Positioning is king in this new invisible funnel.
When the funnel you've been optimizing is no longer the funnel where decisions are made, the playbook changes. In a world where AI sits between you and your customer across millions of conversations daily, being able to influence that conversation becomes just as important as being able to track clicks. Positioning is king in this new invisible funnel.
Positioning is what the market believes about you, where that belief comes from, and whether it's working in your favor when you're not in the room. What AI models say about you is a function of what they've ingested — review sites, Reddit threads, analyst pieces, competitor pages, your own documentation. Some of it's accurate. Some of it's outdated. Some of it was written by people who don't like you. The model doesn't know the difference, and neither does your buyer.
Brand as Infrastructure
This is what it means to treat brand as infrastructure. Not a rebrand, not a campaign, not a narrative refresh — but the underlying system that determines how your company is understood when no one from your team is present to clarify. Infrastructure in the sense that it needs to be built deliberately, maintained consistently, and owned clearly. Most companies don't have a clean answer to the question: if an AI synthesized everything publicly available about us right now, what would it conclude?
What makes this genuinely interesting from an organizational standpoint is that solving it doesn't live cleanly inside any one function. Marketing owns the messaging but not the documentation. Product owns the docs but not the review sites. Sales knows what objections are coming up but isn't feeding that signal back into how the brand is being shaped upstream. PR influences analyst coverage and press but rarely talks to the team monitoring what AI models are actually saying. These functions have always been connected in theory. In practice, most companies run them as separate pipelines that occasionally sync.
That starts to break down in this environment. If a wrong claim is circulating in AI conversations — something outdated about your pricing, a missing compliance certification, a competitor narrative that's gone unchallenged — the fix might require product to update documentation, marketing to publish something authoritative, and comms to address the source. No single team owns that loop. And without someone owning it, it just doesn't get closed.
The companies that adapt to this will likely look different structurally. Brand stops being a marketing subdiscipline and starts functioning more like a cross-functional layer — something closer to how the best companies think about data or security. Not owned by one team, but with clear accountability, shared visibility, and regular input from across the organization. The question of "what is the market being told about us" becomes as operational as "what is our pipeline."
Brand stops being a marketing subdiscipline and starts functioning more like a cross-functional layer.
That's the shift worth paying attention to. Not just that AI is changing how buyers find you, but that it's changing what kind of organizational capability you need to manage your position in the market at all.
Our prediction is that newer companies will have roles that seem like many of the traditional segmented ones rolled into one — a reflection of how interconnected brand, product, and go-to-market have always been, but are now impossible to treat separately.