By Yanisa Chubplun, Integrated Communications Manager, Foundeast
The New Marketing Funnel: How AI Answers Are Reshaping Early-Stage Discovery
01 The invisible first touchpoint
A familiar pattern is beginning to show up in marketing reports: a detailed, direct enquiry from a company that appeared nowhere in the measurable funnel beforehand — no campaign tag, no referral source, and an opening line close to “we’ve narrowed it down to three agencies.” Nobody on the receiving team can place where the company came from.
One possibility now sits outside the usual report: the prospect may have asked an AI assistant about the category, checked a few names against their websites, and arrived with much of the early thinking already done.
02 The marketing funnel still exists — the entrance has moved
Nothing about this makes the marketing funnel obsolete. Awareness, consideration, and decision still happen in roughly that order, and a website visit or an enquiry form still marks the moment a brand can finally see the buyer. What has changed is what happens before that moment becomes visible. A synthesised answer from ChatGPT, Gemini, Perplexity, or Google’s AI Overviews can now sit ahead of the search results, the LinkedIn post, and the website itself, introducing a category, naming a few providers, and shaping which of them feel credible enough to look into further. That’s an additional layer of discovery in front of the channels marketing teams already track, not a replacement for them.
03 Why this matters now
The scale involved is not trivial. Google said in June 2026 that AI Overviews had passed 2.5 billion monthly active users, with AI Mode separately surpassing 1 billion. That doesn’t mean every buyer’s journey starts with an AI answer, but it does mean the odds of a prospect encountering one before finding a company’s own content are no longer negligible.
What happens after that encounter is worth separating from the encounter itself. Pew Research Center’s March 2025 analysis of US browsing behaviour found that when a Google search produced an AI summary, users clicked a traditional search result in 8% of visits, against 15% when no summary appeared. That’s a meaningful drop in visible traffic, though it says more about click behaviour than about lasting impression, and it reflects US behaviour specifically, not a global pattern.
On the B2B side, 6sense’s 2025 Buyer Experience Report found that 94% of surveyed buyers used large language models somewhere in the buying process, particularly to compare offerings and synthesise information. AI is joining the evaluation process, not replacing the conversations and direct validation that follow.
04 The new journey: question to contact
A more accurate route through today’s marketing funnel might run: question, AI answer, brand impression, verification, shortlist, contact. A buyer asks about a problem or category. The AI synthesises an answer that introduces ideas, criteria, and names. The buyer forms an early view of who looks credible, then verifies that impression against websites, media coverage, executive profiles, and reviews. A shortlist forms. Contact happens once most of the evaluation is already done.
Picture a European retailer weighing market-entry support in Thailand. Before anyone at a candidate agency knows the enquiry exists, the buyer has likely asked an AI tool what such an agency typically does, noted a few names it mentioned, and gone looking for evidence those names hold up — recent client work, press coverage, a founder who writes credibly about the market. By the time an email arrives, the agency is already being checked against an impression it never shaped directly.
05 What marketing teams may be missing
Most teams can see website visits, organic traffic, form fills, and conversions. What they generally cannot see is the AI conversation that happened before any of that: the question asked, the answer given, the names mentioned or left out. That’s not a reason to assume every direct visit began with an AI prompt. It is a reason to treat conventional attribution as a partial account of how buyers actually arrive.
06 What brands need to build
The practical response isn’t chasing a mention in a particular AI answer, since nobody can promise that outcome. It’s building a body of information that holds up wherever a buyer’s research starts:
- Consistent positioning across the website, press coverage, and executive profiles
- Original content answering real questions rather than restating service pages
- Credible third-party coverage
- Visible executive expertise
- Content structured clearly enough for readers and retrieval systems alike
- Periodic checks on how AI tools currently describe the company and its category
None of that guarantees a recommendation — it means that when a buyer checks, what they find agrees with what they were told.
07 A better way to think about acquisition
“The task is not simply to appear in an AI answer. It is to make sure the evidence a buyer finds afterwards confirms rather than contradicts that first impression.”
Acquisition, in this version of the marketing funnel, can begin before any click a company would ever see. The task is not simply to appear in an AI answer. It is to make sure the evidence a buyer finds afterwards — the website, the coverage, the people behind the brand — confirms rather than contradicts that first impression.
08 Where Foundeast comes in
This is the kind of gap Foundeast spends time on with clients: mapping where a buyer’s research actually starts, checking how AI tools currently describe a brand against its competitors, and tightening the positioning, content, and coverage behind that first impression. If you’re not sure what your category looks like from inside an AI answer, that’s usually a reasonable place to start a conversation.