An infographic visualizing that 94% of B2B buyers use Large Language Models (LLMs) and showing usage peaks during the comparison and synthesis stage of the purchasing funnel.
| |

94% of B2B Buyers Now Use LLMs to Research Software — Is Your Company Visible When They Ask?

ByJohn Mecke

B2B buyers are using LLMs to research enterprise software — and most SaaS companies have no idea they’re being evaluated. If you run a pre-seed or early-stage B2B SaaS company, you are probably spending most of your marketing budget trying to rank on Google. You invest in SEO, pay for sponsored placement on G2 and Capterra, and optimize your website for search. That strategy made sense in 2022. In 2026, it is dangerously incomplete.

A seismic shift has occurred in how enterprise software buyers conduct research — and it has happened faster than most vendors realize. Two years ago, 68% of buyers told TrustRadius that generative AI had no impact on their buying process. Today, according to 6sense’s 2025 global study of nearly 4,000 B2B buyers, 94% use large language models during the purchase journey. That is not a gradual trend. That is a market transformation.

This post breaks down what the latest research tells us about how and when enterprise buyers use LLMs like ChatGPT, Claude, and Gemini — and what it means for how you build your go-to-market strategy as an early-stage SaaS founder.

The 94% Number — And Why It Needs Unpacking

Let’s start with the headline. In 6sense’s 2025 Buyer Experience Report — based on nearly 4,000 global B2B buyers — 94% reported using LLMs at some point during a software purchase. That number is remarkable. But it is also somewhat misleading if taken at face value, because it obscures a more nuanced story.

The research reveals that LLM use does not peak at the start of the buying journey — it peaks in the middle. Buyers are not primarily asking ChatGPT “who are the top vendors for X software?” as their first move. Instead, they use LLMs after they have already identified a shortlist — to compare offerings side-by-side, synthesize vendor documentation, model costs, draft RFP language, and build implementation plans.

Key Stat: 94% of B2B buyers use LLMs during their purchase journey — but peak usage is mid-funnel, during vendor comparison and synthesis, not at initial discovery. Two separate visibility problems must be solved.

This distinction matters enormously. There are effectively two separate problems to solve: being surfaced during initial vendor discovery queries, and being accurately represented when a buyer in active evaluation asks an LLM to compare you against your competitors. Both require investment, and they require different approaches.

One in Three Buyers Now Starts Research in an LLM — Not Google

So if LLMs are primarily a mid-funnel tool, is your SEO strategy still safe? Not quite. A separate study by Responsive — based on more than 350 B2B buyers worldwide — found that 32% of buyers now use generative AI as much as or more than traditional search engines when researching vendors. Two-thirds say they rely on AI chatbots at least as much as Google or Bing across their evaluation process.

In the technology and software category specifically — where your buyers almost certainly sit — that number jumps to 80%. Four out of five enterprise tech buyers say they use AI tools at least as much as traditional search when evaluating vendors. If your company is not showing up accurately in LLM-generated responses, you are invisible to the majority of your potential buyers during evaluation.

There is also a generational dimension that makes this trajectory irreversible. According to Leadscale’s 2026 analysis of multiple buyer studies, 85% of buyers aged 25–34 use AI for supplier research — compared to just 23% of buyers aged 55–64. The younger cohort is not future buyers. They are today’s procurement leads, project managers, and evaluation committee members. This is not an early-adopter curve you can afford to wait out.

The Dark Funnel Has a New Address: Inside the LLM

Here is where the problem gets existential for early-stage companies. According to analysis by Spotlight Analyst Relations and Profound, more than 20 million prompts per day on ChatGPT alone are tied to B2B buying decisions. When you include activity across Claude, Copilot, and Perplexity, that figure exceeds 80–100 million B2B research prompts per day. None of that activity shows up in your website analytics. None of it generates a lead notification. It happens entirely inside the LLM interface — the new dark funnel.

The traditional dark funnel — buyers researching you without your knowledge — was already a challenge for early-stage founders. The LLM dark funnel is orders of magnitude larger. And unlike a Google search that might eventually result in a click to your website, an LLM interaction that does not mention your company generates zero downstream signal whatsoever. You never know you lost.

“If ChatGPT does not mention your company, your buyer may never encounter you during their research. The dark funnel expands: research happens inside AI agents, not inside your analytics.” — Leadscale, 2026

And the distribution of visibility within LLM responses is brutally concentrated. Research from BrightEdge and Amsive shows that AI platforms cite only 3–4 brands per response on average, with the top 20 domains capturing 66% of all AI citations. This is a winner-takes-all dynamic. There is no page two in an LLM response.

For founders thinking about competitive positioning and eventual exit value, this concentration matters enormously. LLM citation authority is becoming a component of brand equity — one that sophisticated acquirers and investors will increasingly scrutinize. See our analysis of how enterprise SaaS win/loss dynamics are shifting in 2025 for the downstream deal implications.

What LLMs Actually Trust: The Citation Hierarchy

Understanding why certain companies get cited — and others do not — requires understanding how LLMs construct answers about vendor landscapes. They do not simply surface the best-funded companies or those with the most backlinks. They synthesize from sources they have been trained on and, for models with real-time web access, from content they can actively retrieve and verify.

6sense research reveals a striking downstream effect: when buyers ask ChatGPT or Claude for software recommendations, the models frequently reference G2, Capterra, and TrustRadius as sources — treating aggregated peer reviews as more authoritative than vendor marketing claims. This creates a reinforcing loop: your G2 profile and review volume directly influence whether LLMs cite you, which influences whether buyers encounter you during LLM-mediated research.

Notice what is absent from the high-trust tier: gated whitepapers, login-protected case studies, and vendor blog posts that lack external trust signals. If the majority of your content is locked behind a form fill, LLMs cannot access it, cannot synthesize it, and cannot cite it on your behalf. Your content infrastructure is either working for you in the LLM channel — or it is not working at all.

The Buying Journey Has Changed — But the Fundamentals Have Not

It would be easy to read all of this and conclude that the LLM shift has upended everything about enterprise software sales. But the 6sense data adds an important corrective. Despite the explosion in LLM usage, buyers still average 16 interactions per person with the winning vendor — essentially unchanged from 2023. LLM research has accelerated qualification and comparison, but it has not replaced the human conversations that ultimately close deals.

What has changed is the quality and depth of the preparation buyers bring to those conversations. According to Responsive’s study, 90% of buyers conduct extensive research before making first contact. When a buyer finally reaches out to your sales team, they have already formed strong views about your positioning, your differentiators, and how you compare to two or three competitors — views shaped largely by what LLMs told them.

Key Insight: The pre-contact favorite still wins 80% of deals. LLMs are now the primary mechanism that determines who makes the shortlist before the first conversation ever happens.

The 6sense research also reveals something counterintuitive: 58% of buyers contacted vendors earlier than usual — specifically to ask questions about AI capabilities that LLMs could not reliably answer. The models accelerate the pre-sales phase while creating new categories of questions that only a live conversation can resolve. For a deeper look at how buyer behavior plays out in actual deal outcomes, see our analysis of why enterprise SaaS deals are won and lost.

What This Means for Early-Stage B2B SaaS Founders

For a pre-seed or seed-stage SaaS company, the LLM shift creates both a threat and an asymmetric opportunity. The threat is obvious: if you are not represented in LLM responses, you will not make shortlists, and you will not know you are missing. The opportunity is less obvious but potentially significant: LLMs create a new pathway to credibility that does not require a massive marketing budget or years of category presence. For context on what the broader funding and competitive landscape looks like for early-stage companies today, see our State of Seed 2025 analysis.

Large incumbents with outdated, gated content libraries may actually fare worse in LLM citations than nimble early-stage companies that publish structured, ungated, original research. LLMs reward epistemic authority — the ability to demonstrate domain expertise through content that third parties can find, verify, and synthesize — not just brand recognition or ad spend.

Four Practical Priorities for LLM Visibility

Based on the research reviewed here, early-stage SaaS founders should focus on the following four actions:

  • Publish original primary research. LLMs cite authoritative, structured content from credible sources. A well-executed primary research study — even a 250-respondent panel establishing market benchmarks — gives you the kind of citable, quotable, third-party-verifiable content that LLMs pull from. This is no longer a nice-to-have for category leaders. It is a table-stakes visibility play.
  • Build your G2 and TrustRadius profile deliberately. LLMs treat peer review aggregators as trust signals. A company with 40 verified G2 reviews and a detailed comparison page is more likely to appear in LLM-generated vendor comparisons than a company with a polished website and no third-party validation. Invest in systematic review generation from your existing customers.
  • Make your best content ungated. Every whitepaper, case study, or research report sitting behind a form fill is invisible to LLMs. If your goal is to be cited in AI-generated responses during buyer research, you need to put your highest-authority content where models can reach it — ideally on your own domain with structured metadata.
  • Create content that answers mid-funnel comparison questions. LLM usage peaks when buyers compare options. Create explicit, detailed content comparing your solution to your top two or three competitors — structured so LLMs can parse and cite it. Do not leave that comparison narrative to chance or to what your competitors publish about themselves. Our competitive analysis playbook provides a framework for building this kind of intelligence asset.

The Window Is Open — But Not Forever

One of the most striking findings from the 10Fold / Sapio Research study of 400 senior marketing executives is that only 11% of B2B brands have the majority of their content AI-discovery ready. That means roughly 89% of your competitors are not yet optimized for LLM visibility. The citation slots that will define LLM-mediated vendor discovery are consolidating right now — and once they consolidate around a set of incumbents with strong AI presence, they will be very difficult to displace.

The companies that move now — publishing original research, building peer review volume, creating ungated comparison content, and structuring their web presence for AI crawlability — will establish citation authority that compounds over time. LLMs reinforce what they have already learned. Early visibility breeds persistent visibility.

Meanwhile, the volume of LLM-based buying activity is only accelerating. Gartner has projected that by 2028, 90% of B2B buying will be agent-intermediated — meaning AI systems will be conducting research, generating shortlists, and in some cases initiating vendor outreach on behalf of human buyers. If your company is not visible and accurately represented in LLM responses today, you are building toward a future where you are structurally absent from the most important channel in enterprise software sales.

For early-stage founders thinking about exit timing, LLM visibility is also becoming a GTM quality signal that sophisticated acquirers will evaluate. The strength of your demand generation infrastructure — including whether your brand appears in AI-generated competitive comparisons — increasingly reflects the durability of your ARR. For more on how acquirers evaluate early-stage SaaS businesses today, see our overview of enterprise value benchmarks for pre-seed and seed acquisitions in 2025.

Conclusion: The Funnel Starts Before You Can See It

The data is unambiguous. Enterprise software buyers are using LLMs — not as a curiosity or a side tool, but as a primary research infrastructure. Ninety-four percent use them during the buying journey. In the tech sector, 80% use them as much as or more than search engines. More than 20 million B2B buying prompts are submitted to ChatGPT every single day. And in each of those responses, 3–4 vendors get cited. The rest are invisible.

For early-stage B2B SaaS companies, this creates a strategic imperative at least as urgent as SEO was in 2015 or content marketing was in 2018. The companies that build LLM visibility now — through original research, peer review presence, ungated authority content, and structured AI-readable web architecture — will have a durable competitive advantage that compounds as LLM-mediated buying becomes universal. The first step is understanding precisely why you are winning and losing deals today. Our AI-accelerated win/loss analysis service was built specifically for seed-to-Series B SaaS companies navigating exactly this environment.

Your buyers are already asking AI about you. The question is whether AI knows who you are.

Ready to Close Your LLM Visibility Gap?DevelopmentCorporate LLC helps pre-seed and early-stage B2B SaaS companies build the research assets, competitive intelligence, and content infrastructure needed to achieve LLM visibility and accelerate demand generation.Learn more at developmentcorporate.com →

Sources and Methodology Notes

Gartner — B2B Buying Behavior Forecast, 2025

Similar Posts