An operations lead is shopping for inventory software that works with the company’s ERP and supports a warehouse team in Mexico. She asks an AI assistant for options, checks the vendors’ integration docs and brings two names to her team. By the time procurement joins the discussion, the team has a list of questions about rollout, support coverage and the cost of moving its data.

That is one way AI in B2B buying can shape a vendor decision. A chatbot might surface a name the team missed, summarize competing proposals or help a buyer prepare for a demo. Each task tells us something different about the role AI plays in a purchase. Those differences matter when reading coverage that puts AI adoption at 89%. The PuroMarketing article that prompted this analysis.

What Forrester’s 89% actually measures

In Forrester’s 2024 Buyers’ Journey Survey, 89% of respondents said they used generative AI in at least one part of their buying process. In January 2026, Forrester reported that the figure was 94% in its 2025 survey. A buyer who uses an assistant to summarize a proposal counts in that measure, even if the team entered the evaluation with a clear favorite. Forrester’s 2024 survey analysis and its January 2026 update.

Gartner looked at a recent purchase. Of 645 B2B buyers surveyed in August and September 2025, 45% said they used generative AI, primarily to gather information about products and vendors. Gartner published the results in May 2026. The figures come from different samples and survey questions. One captures reported use at any point in the process, while the other asks respondents to recall what they used on a recent purchase. Gartner’s survey release.

How AI can influence the vendor shortlist

G2 surveyed 1,076 people who make or influence B2B purchase decisions across North America, EMEA and APAC in March 2026. Among software buyers polled, 69% said an AI chatbot led them to choose a different vendor than they originally planned. 33% said they bought from a vendor they hadn’t heard of before. The comparison is with the buyer’s initial plan, which may have included keeping an incumbent or selecting another new vendor. The survey captures buyers’ own accounts of what influenced them. G2’s 2026 AI Search Insight Report.

6sense’s 2026 research adds context. It found that 81% of surveyed buyers used large language models somewhere in the process, with use concentrated around the middle of the evaluation. Buyers often used them to organize or check information already in hand. The international study focused on business purchases worth at least $25,000 annually. Taken together, the studies illustrate two ways AI can shape a shortlist. An assistant may introduce a vendor or help the buyer assess one that was already on the list. 6sense’s 2026 B2B Buyer Experience Report.

What buyers do after a chatbot recommends a vendor

Semrush surveyed U.S. B2B professionals in March and April 2026 and based its findings on the 519 respondents who already used AI for work. When a chatbot mentions a vendor, 71% of that group said they visit the vendor’s website and 63% said they look up the company on Google. Those answers describe the habits of existing AI users. Semrush’s survey of U.S. B2B professionals.

Gartner also found that 69% of its survey respondents prefer to check AI-generated information with a sales rep. For the inventory software buyer, the vendor’s site can confirm whether a connector supports the right version of the ERP. A conversation with the vendor can clarify who owns the rollout, what happens if a migration goes wrong and whether the quoted price includes support. Gartner’s findings on buyer validation.

Make your offer easier to evaluate

Start with the questions that come up when buyers take a shortlist to procurement or finance. Ask your sales team which details prospects repeatedly request and which claims customers ask them to substantiate. That can help you prioritize the information a buyer should be able to find before a call and bring into the room afterward.

• Spell out the fit. Put supported integrations, service areas, implementation requirements and plan limits where buyers expect to find them. If a connector costs extra or only supports certain versions, explain that alongside the feature.
• Give buyers something to verify. Publish customer stories that explain the starting conditions, what your team did and how the outcome was measured. Keep documentation current and offer references when customers are willing to discuss their experience.
• Check what assistants say about your offer. Try questions a prospect might ask about your category, their location and a specific requirement. Keep a record of the tools, prompts and sources. Use recurring errors to prioritize content fixes, and ask prospects how they heard about you during discovery.

For a vendor serving buyers in the U.S. and Latin America, a practical starting point is a bilingual implementation page covering rollout, support hours and regional availability. A buyer who finds you through an assistant can use the same page to check the recommendation and brief the people who will approve the purchase.