Updated on 7 November 2026, by João Leite, Founder of murmur.

The B2B buyer describes the problem and the company profile to the AI, receives a short list of suppliers and checks the answer. A short list is the small group of suppliers invited to submit a proposal. In the 2026 surveys with software and technology buyers, AI became one of the main sources of that list, alongside review sites. For the company that sells, appearing in ChatGPT at that moment calls for a published answer to each question from the people who approve the purchase.

  • Most software buyers have already asked a chatbot for a recommendation. In G2’s July 2026 survey, 82% had done so in the last two years.
  • For half of them, the biggest effect of AI comes from cutting options. Of those who asked a chatbot for a software recommendation, half told G2 that the biggest effect came from narrowing down and comparing the list.
  • The supplier’s website also shapes the short list. In G2’s March survey, 36% of purchase decision-makers cited the supplier’s website, the same share as market research firms.
  • In the decarbonization sector, who appears on the list changed with the type of company in the question. In a murmur study with 25 questions, the second most cited Brazilian company went from 6 of 15 answers in the general question (40%) to 4 of 21 (19%) in the questions that stated the type of company.

In B2B purchasing, murmur, a GEO agency in Brazil that places brands in the answers of ChatGPT, Claude, Google AIO and other AI agents, starts from the questions the client’s buyer asks AI models. It studies the client’s sector and writes the guides that answer each person on the committee. Then it tracks, week by week, in which questions the company appears in ChatGPT, Claude, Perplexity and Google AIO, the AI summary at the top of Google search.

The B2B buying committee: several people, and a different concern for each one

On the buyer’s side, the B2B decision belongs to a buying committee, the group of people from different areas who approve the supplier. The size of that group, by Forrester’s calculation, is in the guide on how to explain GEO to your boss or the board. The person who will operate the solution wants to know whether it solves the day-to-day problem. The technical area looks at integration with the company’s systems, finance looks at total cost, and procurement looks at contract, deadlines and reference clients. For the company that sells, each of these concerns calls for a published answer, not a generic page for everyone.

How many B2B buyers already use AI to choose a supplier

Most B2B software buyers already use AI to reach suppliers, according to the 2026 surveys with decision-makers. G2, which sells services to software suppliers, published a survey in July with more than a thousand software buyers, without saying which countries they were from. Of these, 82% asked an AI chatbot for a software recommendation in the last two years, and the page does not say which chatbot. The TrustRadius surveys, with technology buyers, and the HubSpot surveys, with CRM purchase decision-makers, are in the guide on how to get your company recommended by AI. None of these surveys reports the result for the Brazilian buyer.

At what point in the purchase the B2B buyer asks AI

AI enters B2B purchasing from the start, and some buyers feel its biggest effect when cutting options. Forrester, an American research firm, wrote in January 2026 that the B2B buyer starts the purchase with generative AI search. In G2’s July survey, half of the buyers who asked a chatbot for a software recommendation said AI had the biggest effect when narrowing down and comparing the options. How to turn the questions from each stage of the purchase into pages is in the GEO guide for B2B companies.

What the B2B buyer writes in the question to build the short list

To build the short list, the B2B buyer writes their company profile into the question: size, sector and problem. In murmur’s decarbonization study, seven of the 25 questions stated the type of company: industry, midsize company, agribusiness, supply chain, logistics, exporter and retail. In the 21 answers to them, the order of the Brazilian suppliers changed. The second most cited Brazilian company in the general question, present in 6 of the 15 answers (40%), appeared in 4 of the 21 (19%). Another one, which was not among the top four in the general question, appeared in 5 of the 21 (24%).

In this sector, in a September 2026 round, who appears on the list changed when the question stated the type of company. The count of how many answers included a Brazilian company, with and without the profile, is in the GEO guide for B2B companies, in the five sectors measured. For sellers, the question the buyer asks including their size and sector is the one worth testing first.

The B2B buyer checks the AI answer before deciding

The B2B buyer uses the AI’s list and checks what it said before deciding. The TrustRadius survey that shows how many check is in the guide on the buyer who trusts the AI recommendation and checks before deciding. In the same survey, the product demo, the free trial, previous experience and reviews from other users are among what weighs most in the choice. For the company that sells, the fact the AI repeated, such as price, delivery time or who it serves, must be the same one the buyer finds when they check.

Beyond AI, the B2B buyer listens to reviews, market research, colleagues and the supplier’s website

Beyond AI, the B2B buyer builds the short list with review sites, research firms, the supplier’s website and colleagues. In March 2026, G2 surveyed 1,076 software purchase decision-makers from North America, Europe, the Middle East, Africa and Asia-Pacific. After AI chatbots and review sites came market research firms and the supplier’s website, with 36% each, colleagues, with 33%, and independent forums, with 30%. The numbers for chatbots and review sites, which swapped places between March and July, are in the guide on the buyer and the AI recommendation.

The weight of each source within AI answers is in three guides. They are the sites that most influence ChatGPT in Brazil, reviews and Reclame Aqui in AI and press relations to appear in ChatGPT.

When the AI gets it wrong or leaves your company off the short list

When the AI gets it wrong about the company or leaves it off the short list, the buyer decides with the wrong information or without that name. The error shows up as a price that has already changed, a service the company never offered or the name swapped for that of a namesake. What Forrester wrote about incomplete information in AI searches is in AI recommendations are not always accurate.

The error is corrected in the sources the AI reads, and the step-by-step guide is in how to correct wrong information ChatGPT says about your company. When the company does not even make the list, the diagnostic, AI by AI, is in why ChatGPT does not recommend your company.

What appearing in ChatGPT asks of sellers

To appear in ChatGPT when the buyer builds the short list, the seller needs to publish the answer to each committee member’s question. This is what is called GEO, the work of making the company appear in the answers from ChatGPT and the other AI models. In B2B, it asks for five things:

  1. One page for each question the approver has: the use, the integration, the cost and the contract, each one with the complete answer, and not one page for each way of writing the same question.
  2. The type of client written down: the size, the sector and the problem, in the buyer’s own words.
  3. The facts in visible text: what the company delivers, the delivery time and the price, or what it depends on.
  4. The same fact everywhere: the price and the scope the same on the website and on the company’s profiles.
  5. Checking the answers: question by question, to see which ones the company appeared in.

The complete step-by-step guide is in how to do GEO for a B2B company.

How to know if your company is on the shortlist that the AI builds

Your company is on the shortlist when its name comes out in the answer to your buyer’s question, with their profile. Repeat each question in ChatGPT, Claude, Perplexity and Google AIO, on different days, because the list changes from one answer to another. Presence is the number of answers in which the brand appears divided by the total number of answers. Separate the questions by person on the committee, and the calculation shows which question the company still fails to answer.

The step-by-step instructions are in how to measure the brand’s presence in ChatGPT, and the guide on the number of questions for measurement covers how many questions to use.

Industry, service and software: the B2B buyer changes the question

The B2B buyer changes the question depending on what they are buying: in industry, they ask for specification and standard; in software, comparison and integration. The full answer, with what murmur measured in B2B service, is in AI recommends industry and service the same way it recommends software. Industry has its own guide, how to do GEO for industries, and so does software, in how to appear in ChatGPT in software comparisons.

At murmur, a specialist studies the questions of each person on the committee

At murmur, the same specialist follows the B2B company from the diagnostic onward and studies the questions that each person on the committee asks the AI models. A committee with many people calls for many answers, and murmur does not set a cap on questions, hours or content. murmur works without a minimum contract term: the client can end it from the first month, and the written content remains entirely theirs. If the AI gets a piece of data wrong that the buyer will check, such as the price or the scope, murmur also helps correct the information.

Mindo, a motion design studio that sells video to companies, is a B2B service case study for murmur. ChatGPT began recommending Mindo in 23% of 134 questions on 13 and 14 July 2026, against 12% in the same questions on 25 and 26 June.

Frequently asked questions

Does the B2B buyer already use ChatGPT to choose a supplier?

The B2B software buyer already asks AI chatbots for supplier recommendations, and most have done so in the last two years. The figure is from G2, which does not say which chatbot.

Does the buyer trust what the AI answers?

The buyer uses the AI to start the list and checks what it answers. The TrustRadius research on this is in the guide on how to make your company recommended by AI.

Does AI replace the review site when building the shortlist?

AI has not taken the place of the review site: in both G2 surveys from 2026, both are among the main sources. The March figure and the July figure are in the buyer and the AI recommendation.

Does my company need a page for each person on the committee?

Your company needs an answer for each question the committee has, because the page that answers the technical person does not answer the finance person. What each page needs to include is in what the B2B product or service page needs to include.

Where to start

Start with the question your buyer would ask the AI, with their size and sector, and see who makes the shortlist. Take that question to ChatGPT, Claude, Perplexity and Google AIO, and count how many answers include your company’s name. murmur’s free diagnostic builds this picture with the questions of the buyer in your market. With the website URL and nothing more, the reading is ready in less than two minutes, with no sign-up. Then, next to each person on the committee, put their question and mark the ones your website already answers.

Read also

References

  • G2, “New G2 Research: AI Is Reshaping How B2B Software Deals Are Won and Lost”, 22 July 2026. https://company.g2.com/news/buyer-behavior-2026 (accessed on 13 October 2026). More than a thousand B2B software buyers and purchase decision-makers, with no country or fieldwork date provided. In the original: “Eighty-two percent of buyers sourced software recommendations from an AI chatbot in the last 24 months” and “half of those buyers said AI had its greatest impact when narrowing and comparing options”. G2 sells services to software vendors.
  • G2, “The Answer Economy: G2’s 2026 AI Search Insight Report”, March 2026. https://learn.g2.com/g2-2026-ai-search-insight-report (accessed on 13 October 2026). 1,076 B2B purchase decision-makers from North America, Europe, the Middle East, Africa and Asia-Pacific. The text lists the sources that form the shortlist after chatbots and review sites: market research firms 36%, vendor website 36%, colleagues 33%, independent forums 30%.
  • HG Insights, “TrustRadius 2026 B2B Buying Disconnect Report Reveals AI Has Changed How Buyers Research, But Not What They Trust”, 15 July 2026. https://hginsights.com/news/trustradius-2026-b2b-buying-disconnect-report-reveals-ai-has-changed-how-buyers-research-but-not-what-they-trust/ (accessed on 13 October 2026). Global study from January 2026 with 1,862 buyers and 444 technology vendors. The text includes the list of the most influential sources in the choice (“product demos, free trials, prior experience, and user reviews”). TrustRadius makes its living from software reviews.
  • HubSpot, “HubSpot AEO data: Buyers using AI search are more likely to purchase”, updated on 15 September 2026, with a January 2026 study of more than 3,000 CRM purchase decision-makers worldwide. https://www.hubspot.com/company-news/aeo-data-buyers-using-ai-search-more-likely-to-purchase (accessed on 13 October 2026). HubSpot sells an AEO product.
  • Forrester, “Forrester’s 2026 Buyer Insights: GenAI Is Upending B2B Buying As Leaders Face Mounting Pressure To Justify Every Dollar Spent”, 21 January 2026. https://www.forrester.com/press-newsroom/forrester-2026-the-state-of-business-buying/ (accessed on 13 October 2026). In the original: “genAI searches are the starting point for B2B buyers”. The page does not state the sample size or the country. Forrester sells research and consulting.
  • murmur, “O que as IAs respondem quando uma empresa procura consultoria de descarbonização”, sector study from September 2026: 25 questions, once in ChatGPT, Claude and Google AIO, 75 answers. Breakdown by type of company: 7 questions and 21 answers; general recommendation: 15 answers. https://murmur.marketing/estudos/descarbonizacao-2026-09/ (accessed on 13 October 2026).
  • murmur, Mindo case study: ChatGPT recommendation in 23% of 134 questions on 13 and 14 July 2026, against 12% on 25 and 26 June 2026. Published with the client’s authorization. https://murmur.marketing/en/cases/mindo/ (accessed on 13 October 2026).
  • murmur, free diagnostic: https://murmur.marketing/en/diagnostic/ (accessed on 13 October 2026).