79% of B2B Buyers Now Use ChatGPT to Research Vendors. Their Bosses Don't Know.

The Secret Research Tool

Seventy-nine percent of global B2B buyers now use AI tools like ChatGPT, Perplexity, and Google AI Overviews as their primary way to research vendors. Not as a supplement to approved processes. As their main method.

Most of their companies have no idea this is happening.

According to a recent report from Improvado, this quiet shift has created what amounts to a shadow AI stack inside enterprises—a parallel research infrastructure that lives entirely outside official procurement channels, governance frameworks, and IT approval processes. Employees paste internal context into consumer AI tools, ask for vendor comparisons, and circulate those outputs as de facto buying recommendations. Legal and security teams often don't realize how much proprietary data is being fed to external models until something goes wrong.

The numbers tell the story of a fundamental rewiring: 57% of all searches now result in zero clicks to any website. Google's AI Overviews appear in 13% of search results, answering queries directly and bypassing traditional vendor sites entirely. The path from "we need a solution" to "here are three options" increasingly never touches a company's official research process.

When the Funnel Disappears

For B2B marketers, this creates an odd problem: they've lost control of the buying journey without necessarily noticing. EndeavorB2B's 2026 trends analysis puts it bluntly—marketers no longer control linear funnels or campaign-driven buyer journeys. Success now depends on designing "the environment, insights, systems, and experiences that empower buyers and algorithms to choose them."

But when the algorithm is ChatGPT making judgment calls about which vendors to surface in a comparison, who actually controls that environment? Not the demand generation team. Not the content strategist optimizing for SEO. The power has shifted to whoever controls the training data, semantic context, and entity mapping in those AI models.

This is happening in parallel with official AI initiatives, creating awkward collisions inside companies. The Content Marketing Institute finds that AI tops the list of areas B2B marketers plan to increase investment in for 2026. Budgets are being allocated. Governance frameworks are being written. ROI dashboards are being built.

Meanwhile, a product manager three floors down is pasting the company's infrastructure diagram into ChatGPT and asking it to recommend monitoring tools.

The Invisible Channel

The practical implications are starting to surface in strange places. On forums like Hacker News, mid-level managers increasingly post variations of the same question: "My team is using ChatGPT for vendor selection. Procurement is furious. Who's right?"

The answer depends on who you ask. From a governance perspective, feeding proprietary context to external AI models is a security risk. From a practical standpoint, it's often the fastest way to get a decent answer. The shadow AI stack exists because it works better than the approved alternative.

This opens up an entirely new category of vendor positioning. Being correctly understood and represented by AI systems becomes as important as ranking in search engines—yet most companies have no dedicated function for "LLM positioning." A few B2B agencies and data providers are already specializing in tuning content and metadata so that Perplexity, ChatGPT, and AI Overviews describe your product the way you want. It's SEO for the post-search era.

The weird part is that 96% of B2B marketers now use AI in some capacity, but the shift isn't just about analytics or automation. It's about agentic AI that executes decisions autonomously—surfacing vendors, making initial comparisons, drafting recommendations. That autonomy amplifies both efficiency and mistakes.

What It Means

The most important B2B technology platform of 2026 might be something your CIO has never approved. That's not a criticism of CIOs—it's a recognition that the pace of informal tool adoption has outrun the pace of formal governance.

For vendors, the implication is uncomfortable: you can have perfect SEO, a beautiful website, and a well-staffed demand gen team, and still never appear in the research process if an AI assistant doesn't surface you in its initial comparison. For buyers, the implication is riskier: the vendor shortlist they're presenting to stakeholders might be based on context and data that was never meant to leave the building.

The shadow AI stack isn't going away. It exists because it's faster, easier, and often more helpful than the official alternative. The question is whether enterprises will recognize what's already happening and build governance around it—or whether the collision between official process and actual practice will keep getting messier until something breaks.