How Buyer Experience Offsets LLM-Induced Traffic Loss

The rise of large language models (LLMs) has sparked new concerns for B2B marketers: if buyers increasingly rely on AI for their solution research, are they still going to visit provider websites?

The early indications are that vendor websites are losing traffic. But does that mean fewer opportunities to influence and convert buyers?

These are valid questions, but it is critical to understand that not all web traffic is created equal. A decline in overall website visits does not necessarily mean providers are missing out on in-market today. As we describe below, our research suggests that the traffic being displaced by LLMs is more likely to emanate from the professionally curious — future buyers, competitors, students, or practitioners researching for reasons other than an active purchase cycle. By contrast, in-market buyers continue to engage directly with vendors at a rate consistent with previous years.

Understanding this difference is critical. While the loss of curiosity-driven traffic may affect surface-level metrics like impressions or clicks, it does not appear to be impacting the real work of advancing deals. Buyers evaluating investments valued in the tens or hundreds of thousands or more are still showing up, still seeking validation from vendor content, and still engaging in a substantial number of interactions — both human and digital.

This report argues that buyer experience — not raw traffic volume — is the real hedge against LLM-induced disruption. Vendors that deliver clear, credible, and confidence-building interactions ensure that when real buyers are in market, they find what they need and remember it.

Who’s Really Gone Missing (And Who Hasn’t)?

At present, the decline in web traffic driven by LLM usage is unlikely to represent in-market buyers. Our most recent round of global research across nearly 4,000 B2B buyers showed clearly that when buyers are actively evaluating a purchase, they engage directly with vendors at roughly the same level as in prior years. Instead, the traffic that appears to be “lost” is more plausibly coming from future buyers – individuals who have a personal or professional interest in the provider’s content and solutions, but who are not part of an active buying group. For these individuals, the summarized and synthesized content provided by LLMs is more likely to satisfy the moment’s information needs.

In contrast, there is virtually no chance that active, in-market buying group members will rely solely on LLMs for information about providers they are evaluating.

The Facts

Nearly all buyers — 94% — said they used LLMs at some point in their buying journey.
However, buyers continue to report roughly the same number of interactions per person, per vendor. This year, the average was 16 interactions with the winning vendor — one fewer than in 2024 and statistically the same as in 2023.
Buyers still rely on outside resources (analysts, consultants, investors, etc.) at levels equal to, or in some cases higher than, before.
Whether buyers use LLMs in their journey or not explains less than 2% of the variability in:
- Buying group size
- The point of first contact with sellers
- The number of vendors evaluated
- The length of the buying cycle
- The number of interactions between buyer and provider

So far, LLMs are not taking the place of humans in the buying group, not pushing sellers or outside experts out of the process, not shortening buying cycles, and not eliminating vendors from shortlists. In fact, the average buying group evaluated more vendors this year (5.1) than in prior years (4.5).

How Buyers Are Using LLMs

Active buyers continue to engage in the traditional activities of the buying journey at the same volume. Yet with nearly all adopting LLMs, it begs the question: what, exactly, are they using them to do?

LLM usage appears to be additive. The heaviest use of LLMs comes in the middle of the buying journey, with 57% of buyers in our study saying they used them at that stage. Forty-five percent said they used them at the beginning, and usage drops off sharply in the later stages, with 15% or fewer reporting LLM use near the end.

On average, buyers reported using LLMs for two broad categories of use cases: 1) synthesizing and summarizing information about vendors, and 2) generating net new content to aid in acquisition and implementation planning.

In the chart below, the most common use case is comparing vendor offerings side-by-side, which was statistically more common than all others. Just below that, six use cases were selected at statistically equivalent rates. Most of these mid-level use cases (apart from forecasting implementation timelines) involve summarizing and synthesizing the large volumes of content buyers consume during their journeys — whether that’s internal stakeholder input, third-party reviews, or vendor proposals.

By contrast, tasks such as forecasting implementation, simulating total cost of ownership (TCO), and creating RFPs require LLMs to generate net new content. These content-creation use cases were the least common, with simulating TCOs and creating RFPs statistically less common than all others.

It appears that right now, the primary role of LLMs is to help buyers synthesize the large volumes of content they consume during their journeys. This lines up with our broader finding that traditional activities are still being carried out at the same levels as before. LLMs are augmenting the process, not radically transforming it.

Not Yet, But Never Say Never

That doesn’t mean change won’t come. Just over half (51%) of our respondents foresee AI augmenting their use of analysts or consultants, while 28% expect AI to replace them. About 20% said they anticipate no change. With roughly 80% expecting at least some impact on how they hire (or don’t hire) outside help, it will be important to track how this influence evolves in future waves of our research.

Even at the world’s largest companies, purchases that cost anywhere from tens of thousands to millions of dollars represent significant investments and personal risks for the people making them. Even if LLMs were fully trustworthy today, most buyers would still want to validate the details themselves before moving forward.

Future Buyers Opt for LLMs

While LLMs aren’t displacing vendor content for active, in-market buyers, they may still affect brand visibility among those not yet in-market. Buyers who once might have visited vendor sites out of curiosity — students, practitioners, or future decision-makers — now have their questions answered by an LLM instead. This shift means that these off-cycle information consumers may miss a brand’s repositioning, rebranding, messaging, or portfolio enhancements. This should concern marketers, since, as we describe below, buyers fill most slots on their shortlists the moment they go into market. They may be creating those shortlists prior to engaging buyer content.

How Buyer Experience Can Mitigate LLM Traffic Loss

With this at stake, buyers’ prior experience becomes even more crucial. Our research reveals that nearly all buyers (97%) know at least one vendor on the shortlist. Moreover, the average buyer reports having prior experience with 75% of vendors on the Day One shortlist, an increase from 69% last year. Consequently, the buying group successfully populates the Day One shortlist with nearly four out of the five vendors they will evaluate, leaving only one spot open for a vendor the buyer has not had prior experience with. Notably, buyers purchase from the Day One shortlist 95% of the time.

Our data also indicate that today’s buyers overwhelmingly (85%) purchase from brands with which they had personal experience prior to their current buying journey. In essence, brands must earn an evaluation today to remain competitive in future evaluations. Every buyer interaction must serve two purposes: advancing current deals while leaving a lasting impression that can influence future cycles.

Clear, credible, and confidence-building experiences ensure that when future buyers eventually enter the market, they remember who you are — even if their early questions were answered elsewhere.

What the LLM-traffic Loss Does Mean

LLMs are here, and buyers are using them. But LLMs are not rewiring the buying journey — at least not yet. B2B purchases, which often represent significant investments for companies and significant risks for the business-people making the purchase decisions, still require substantial hands-on due diligence and evaluation, multiple perspectives, and repeated interactions with vendors.

For marketers, the path forward is less about chasing traffic and more about ensuring that every experience — human or digital — builds the trust and memories that endure.

And while buyers are still using vendor content and trusted third-parties during their buying journeys, it would appear that they are substantially less likely to visit vendor sites during periods when they are not in market. Where in the past, a repeat buyer in a given category might visit relevant vendor websites from time to time as a way of staying up-to-date, those casual information consumers are more likely to query and LLM and forego visits to vendor websites.

Because we know that buyers set their shortlists early, this suggests that shortlists are now being made by buying groups whose most recent impressions of brands come largely from LLMs. It is incumbent on all brands to ensure that they are being represented fully and accurately in LLMs to ensure that out-of-market buyers are consuming accurate and complete information.

In collaboration with 2X

AI is changing how people research, not how they decide. The companies that are winning build credibility into every touchpoint. 2X helps them do it using data, intent, and scalable execution to turn trust into measurable growth.

Keep Reading

Kerry Cunningham and Sara Boostani

Meet the Author