B2B vendor research has moved inside AI tools. Here's what that means for your sales funnel.

Ninety-four percent of B2B buyers now use large language models during their purchasing process, and 83% define their requirements before they ever speak with a salesperson. Those figures, from 6sense's 2025 Buyer Experience Report, frame a shift that is already reshaping how vendors get discovered, compared, and shortlisted.

The mechanism is straightforward. A procurement manager opens Microsoft Word, asks Copilot to draft a vendor comparison for ERP platforms, and receives a synthesized answer drawn from web data, Microsoft Graph signals, and the model's training. The browser never opens. The traditional SEO-optimized landing page never loads.

The infrastructure is already deployed at scale

Microsoft disclosed in its Q2 FY2026 earnings call that Microsoft 365 Copilot had reached 15 million paid enterprise seats, a 160% year-over-year increase. CEO Satya Nadella noted on the same call that daily active Copilot users grew 10x year-over-year and that average conversations per user doubled. Publicis, Fiserv, NASA, and Westpac are among the organizations that have deployed Copilot at scale, according to WebFX's reporting.

Google moved on a parallel track. On January 8, 2026, the company announced that Gmail was entering what it called the Gemini era, embedding AI features directly into the inbox for more than 3 billion users. Gemini now generates synthesized answers when users search their inbox, and a February 2026 rollout brought a personalized "Help Me Write" feature that drafts email responses by pulling context from past threads and Google Drive files. According to Fortune's coverage of the announcement, 70% of enterprise users who used Help Me Write accepted Gemini's suggestion, and 85% said they wanted more personalized AI in Gmail.

Multiple data sources confirm the behavioral shift

The platform adoption numbers track with buyer behavior data from several research firms. Responsive's 2025 buyer intelligence study found that nearly two-thirds of B2B buyers use generative AI as much as or more than traditional search when researching vendors. For technology buyers specifically, that figure rises to 80%, according to WebFX's analysis of the data. G2's 2025 CMO Buyer Behavior Report found that 79% of software buyers said AI search had changed how they conduct research. TrustRadius reported that 72% of buyers encountered Google's AI Overviews during their research, and 90% of those clicked through to cited sources.

McKinsey's data adds a broader consumer layer: shopping-related queries are now the third most popular application of generative AI overall, according to a McKinsey survey cited in Harvard Business Review. Around half of consumers use AI to research products and services, per a Semrush study cited in the same piece, spanning categories from consumer goods to financial services.

AI-generated RFQs are creating a qualification problem for suppliers

Harvard Business Review, in a July 2026 analysis by Graham Kenny and Ganna Pogrebna, identifies a less obvious consequence of AI-assisted buying: a surge in low-signal inbound demand. AI tools now let buyers generate supplier lists, draft procurement emails, and fire polished requests for quotation to a dozen firms in minutes. The result is more RFQs landing in supplier inboxes without more genuine opportunities behind them.

The problem is most acute for specialist firms. Precision manufacturers that advise on materials and tolerances, software firms that configure systems around a client's specific workflow, and engineering consultancies whose value lies in interpreting a brief before proposing a solution all face the same risk: sales teams spend more time quoting and less time diagnosing, advising, and serving real clients. The HBR analysis draws on case studies of small and mid-sized businesses navigating exactly this dynamic in 2026.

Competitive advantage is shifting from direct customer understanding to managing AI-shaped interactions., Graham Kenny and Ganna Pogrebna, Harvard Business Review, July 2026

What this means for your team

• Audit how your brand appears inside AI-generated vendor comparisons. Ask Copilot or Gemini to draft a comparison in your category and assess whether your company surfaces, and how it is described. Structured, authoritative content is more likely to be cited.
• Tighten RFQ qualification workflows. As AI tools make it trivially easy for buyers to blast inquiries, sales teams need tighter qualification criteria upfront, particularly for specialist or consultative offerings where diagnosing the client's problem precedes any accurate quote.
• Evaluate whether your content strategy addresses AI retrieval, not just search ranking. Gemini's inbox integration and Copilot's Microsoft Graph access mean AI models are pulling from a buyer's existing documents and emails as well as the open web. Content that has been referenced, shared, or filed by prospects has a path into AI-generated responses that SEO alone cannot guarantee.
• Track AI-driven pipeline separately. If buyers are arriving with AI-pre-formed vendor shortlists, first-touch attribution and top-of-funnel metrics may no longer reflect where the real decision was made.