These trends have created a lot of speculation on social media and have more than a few revenue teams asking whether the funnel still matters. In other words: Should you shift to straight AI optimization and move on? Or, should you keep your funnels, too, including SEO and all the rest that comes with them?

The answer to this question is one that successful revenue leaders (marketers in particular deserve props for fighting these battles) have been giving to C-suite leaders who want to curtail certain activities in favor of others (e.g., trading brand building in favor of lead-gen) throughout the digital age: “It’s not either-or. It’s both.”

If you feel seen as you read those last two words, we’ve got you, including data to back you up. (Fist bump. OK, stop with the ring across the knuckles. It was funny the first hundred times.)

If you or your team have been struggling with this question, it’s understandable. AI has upended revenue motions in ways that are predictable (AI search) and unpredictable (near-universal adoption of AI by buyers). The short answer is that you still need your funnels, but each page needs to make a complete sales argument so your content serves your human and digital audiences.

Here's why that approach works and how to put it into practice.

Emerging shorthand on the traditional marketing funnel is that AI broke it. Our take? It’s just fractured.

Let me explain: Consider all the pieces of the owned funnel, including the touchpoints and assets buyers use to move from problem to purchase, such as product pages, case studies, testimonials, etc. The pieces still work together, just not always in the order they used to, and every vital piece needs to be able to do the entire job on its own.

The touchpoints may come out of order, but the buying stages haven't changed. Buyers still realize they have a need or problem, then they research solutions, compare vendors, vet claims and choose which vendor, product or service to use.

What’s changed is where those stages happen and who controls the sequence. Research, comparison and vetting used to stretch over weeks as buyers scoured your website, review sites, social sites, forums, analyst reports, sales calls and so on.

Now, a single AI conversation can take a buyer from the starting point (e.g., “what should I be looking for?”) to a shortlist of vendors without visiting your website at all. Forrester found buyers now rank generative AI as their most meaningful source of information, ahead of vendor websites, product experts and sales.

In our 2026 B2B AI Buyer Behavior Survey, 56 percent of B2B buyers said they use AI to learn what matters when choosing a provider. AI often runs this buying-criteria-building stage before you know the buyer exists, in yet another example of the “dark funnel becoming darker.”

Buyers also use AI to check you against those criteria:

We’ve covered that vetting in depth in recent blogs, including AI's New Ad Tax: “What Happens After the Click” and “What Your CRM Cannot Tell You About Buyer Use of AI,” so we won’t rehash it here. The short version is that every piece of the fractured funnel now has AI somewhere in it.

But that doesn’t mean the funnel stopped working. Forrester found that answer engines often deliver incomplete or unreliable information, so buyers go looking for validation from trusted sources such as peers, product experts and industry analysts. AIRI’s own data found that buyers check your claims with AI.

But that’s not necessarily a one-way street; buyers check AI, too. Ultimately, buyers are trying to make the best possible decision. They even check AI with other AI (86 percent of the buyers AIRI surveyed do so at least sometimes), and plenty check AI’s work themselves. As one manager at a professional services firm told us, “We always double-check all the results. We never let AI make the actual decision.” Variations of this theme appeared repeatedly in respondent answers to open-ended questions.

So, the job the funnel was built to do (i.e., moving a buyer from interest to a decision) still gets done. It just happens in pieces, with some of them happening inside ChatGPT, Gemini, Claude or another AI assistant rather than your funnel, and any one of those pieces might be the only one a buyer sees (e.g., a buyer who clicks straight to your pricing page from an AI citation and never sees the rest of your site). Your job is to make sure each piece can carry the whole sale when it needs to.

As we covered in our blog, “AI's New Ad Tax: What Happens After the Click,” a human on your site clicking around and learning about you is a direct line, and that’s still potent. Arguably, site visitors are now more potent because they’re further along in the buying process since they’ve probably already been through AI’s screening and are still interested.

HubSpot’s experience backs that up. HubSpot CMO Kipp Bodnar told Search Engine Journal that “much of that [lost] traffic was never going to turn into revenue,” pointing to old posts on topics like famous quotes and the shrug emoji. Bodnar also said search volume grew on keywords tied to specific products and problems. Notably, HubSpot’s revenue continued to climb even as traffic dropped.

HubSpot’s buyer research lines up. In a January 2026 survey of more than 3,000 CRM decision-makers, 42 percent used AI search in their evaluation, and those buyers were 36 percent more likely to purchase than those who didn’t. So, a late-stage visitor is likely validating rather than browsing.

Some buyers will never visit your site. Others may see only one page. And some will still move through your funnel for deeper validation. This means you need to build for AI assistants and agents, one-page visitors and full-funnel prospects.

In a fractured funnel, you can't assume buyers or their AI assistants have seen anything else you've published.

At AIRI, we say that each key revenue page needs to make a complete sales argument. What is a key revenue page? The term refers to any page buyers or their AI assistants rely on to decide whether you make the shortlist (e.g., product and service pages, pricing pages, comparison pages, case studies, etc.). We borrowed the term from our sister firm BuzzTheory’s AI RevX practice, which evaluates and rates key revenue pages based on their revenue relevance and AI-readiness.

Consider how buyers encounter these key revenue pages. A buyer might click through to your pricing page from an AI citation. Or their AI might pull and summarize a product page without the buyer ever visiting your site. Either way, that page may be your only opportunity to address the buyer, so it needs to say what you do, who it’s for, why it works and why you’re better than the alternatives.

And, most importantly, whenever possible, you must back it up with proof, such as customer results with numbers, case studies, third-party reviews, analyst mentions, certifications, media coverage and other sources of validation.

That said, this doesn’t mean you should scrap the funnel. A well-built funnel still sequences the experience for the prospects who do move through it. Your core claims and headline proof should be mentioned on every key revenue page. The deeper detail behind them, such as the full case study, technical specs or security documents, can live in one place those pages link to. For example, if a customer cut onboarding time by 40 percent, say so on every product and pricing page, and link to the full case study.

Pages built in this way pull double duty. Specific, verifiable claims stated consistently across pages are what AI needs to describe you accurately. Built right, the pages that serve the human visitor also serve your answer engine optimization (AEO), which you need to do anyway.

So the problem isn’t really about choosing between the funnel and AI. It’s about engineering what each piece of content is designed to address and when: