Forrester predicts B2B web traffic could fall by between 50% and 75% over the next 12 to 18 months as buyers increasingly get answers directly from AI instead of clicking through to websites. Speakers at Forrester's B2B Forum said the shift demands a complete rethink of how marketers measure and manage search.

Nicky Briggs, vice president and principal analyst at Forrester, described the challenge created by the move from search engines to LLM "answer engines" as a "visibility vacuum". In 2024, two billion B2B searches took place online every weekday, and 40% resulted in engagement through clicks, according to Forrester's data. Today, search volumes have increased but engagement has fallen.

"Quantity down, quality up. Maybe this will balance out; maybe we'll get away with it," Briggs said. "But we're not in the business of getting away with it. The fact remains that visibility has been cut in half."

There is one upside in the numbers. Forrester found that traffic referred by LLMs is three to four times more likely to convert.

AI search is not a separate discipline

Gerald Murphy, search, AI and consumer behaviour specialist who holds the title 'evangelist' at Semrush, cited company data showing 90% of people now use an LLM alongside a traditional search engine such as Google, while one in three searches carried out in ChatGPT uses SEO data. That means AI search cannot be treated as a separate discipline from SEO, content, paid search and PR. Marketers need a unified approach instead.

Murphy argued that SEO teams must move beyond keywords and think about semantics. When a buyer enters a prompt, AI breaks it into multiple related searches, or "query fanouts", potentially introducing brands and competitors that were not in the original question. SEO teams should analyse these fanouts and share the findings with analytics, content and paid search teams to understand what buyers want and how the brand appears.

"One question in AI can actually impact multiple teams, and we just need to make sure that these teams are sharing information on a regular basis," he said.

Legacy metrics no longer fit

Rankings, clicks and referral traffic lose their value when AI answers a buyer's question without a site visit. Murphy called this out directly.

"The other elephant in the room is that we are actually applying legacy metrics to what is a completely different new system," he said. "Number one position is no longer the goal, particularly when it comes to AI."

Marketers instead need to know whether their brand appears in an answer, where it appears, which competitors get mentioned and which sources get cited. Authority must also be built beyond the brand's own website, because AI systems pull from multiple third-party sources to construct answers.

"Don't chase yesterday's citation," Murphy said. "We need to make sure that we're optimising for relevant articles."

He pointed to Reddit as a platform B2B marketers should watch, given its growing prominence in searches around tool discovery, strategy research and agency and vendor research. "Whenever you look at the trends of Google United States for Reddit in relation to B2B, it's increasing massively. It now ranks for thousands of keywords in the number one position," he explained.

Briggs argued brands should also look internally for authority. "When we can identify the topics that they get passionate about that line up with what we need to be famous for, that's when that authenticity comes through, and the content pretty much writes itself," she said.

Breaking down silos

For Murphy, the changes ultimately demand a joined-up operating model across marketing, spanning content, PR, community and AI optimisation. AI could finally force businesses to adopt the "holistic search strategies" marketers have discussed for years.

"The machine doesn't respect org charts," Murphy said. "The North Star is building an operating system. We need to make sure that both journeys are being looked at from the machine and the human."