Updated July 2026. By Remy Beaumont.
Key takeaways
94% of business buyers now use AI in their buying process, up from 89% a year earlier, according to Forrester's Buyers' Journey Survey, 2025. The AI vendor shortlist is already the default, not an emerging trend.
51% of B2B software buyers now start their research in an AI chatbot rather than a search engine, and 69% ended up choosing a different vendor than they originally planned because of AI guidance, per G2's 2026 research.
One in three G2 respondents bought from a vendor they had never heard of before an AI tool surfaced it. Being unknown is no longer a fixed disadvantage, and being known is no longer a moat.
Trust has not moved to the machine wholesale: Gartner found 69% of buyers validate AI generated insights with a sales rep, because 51% expect AI to serve them misleading information.
The practical conclusion: answer engines learn from third party corroboration, so the coverage, reviews and chatter generated at launch are now training data. Your launch narrative decides your AI shortlist position months later.
Why does this matter now?
Because the person deciding whether your startup gets on the AI vendor shortlist is under 45, and they no longer begin at Google or your website. Millennials and Gen Z became the majority of B2B buyers back in 2022, at 64% of purchase influencers, and Forrester measured the share climbing to 71% a year later (Forrester, via Demand Gen Report). These buyers grew up with consumer software and carry consumer habits into procurement: self serve research, peer validation, and now conversational AI as the first port of call.
The shift happened fast. In January 2025 Forrester reported that 89% of business buyers used AI in the buying process. Twelve months later that figure was 94%, and twice as many buyers named generative AI or conversational search as a more meaningful source of information than any other source, ahead of vendor websites, product experts and sales teams. This is what Forrester now calls zero click buying. If your growth model still assumes the buyer lands on your site early in the journey, it is a model of a market that no longer exists.
What the research shows
Three independent research houses published converging data over the past six months.
Forrester, Buyers' Journey Survey 2025 and The State Of Business Buying 2026. 94% of buyers use AI in the buying process, and 61% use private AI tools supplied by their own organisation. Business buyers are twice as likely as consumers to use ChatGPT and four times as likely to use Microsoft Copilot. Meanwhile the typical buying decision now involves 13 internal stakeholders and nine external influencers, procurement acts as a decision maker in 53% of buying cycles, and more than 60% of buyers use a trial to de-risk the purchase, rising to 78% on purchases of $10 million or more.
G2, The Answer Economy, April 2026. In a global survey of over 1,000 B2B software decision makers, 51% now start research in an AI chatbot, 71% use one at some point in the process, 69% switched away from the vendor they first intended to buy because of AI guidance, and 83% say AI makes them more confident in the final decision. A third bought from a vendor they had never heard of.
Gartner, survey of 645 B2B buyers, published May 2026. Buyers used an average of seven information sources in a recent purchase and 45% used generative AI, mostly to gather vendor and product information. Crucially, 69% validate what the AI tells them with a sales rep, and 51% say they are more likely to encounter misleading information from generative AI than from a human seller. Gartner separately found 67% of buyers prefer a rep free experience, yet also predicts that by 2030, 75% of buyers will prefer sales experiences that prioritise human interaction over AI. Buyers want the machine for research and the human for reassurance.
Why do younger buyers trust an AI answer more than your website?
Because the AI answer feels like consensus and your website feels like a claim. A millennial buyer treats a ChatGPT recommendation the way they treat a heavily upvoted Reddit thread: not as truth, but as a socially filtered starting point that strips out vendor spin. Your homepage says what you want them to believe. The answer engine appears to summarise what everyone else believes. That is classic social proof psychology transplanted into procurement, the same consumer instinct we unpacked in our post on the myth of the rational B2B buyer.
The Gartner validation finding shows the mechanism precisely. Buyers do not blindly trust the machine, 51% expect it to mislead them. They use AI to compress the messy early research phase, then spend their scarce human interactions checking the AI's homework. The shortlist is formed by the machine and confirmed by people. If you are not in the machine's answer, you never reach the confirmation stage.
What does an answer engine actually put on the shortlist?
Corroborated entities. Answer engines weight independent, third party mentions far more heavily than anything you publish yourself: press coverage, review platforms, analyst commentary, community threads, comparison articles. This is why G2 found a third of buyers purchasing from vendors they had never heard of. The vendor was not famous, it was well corroborated in the sources the model draws on.
The market has already repriced this. Profound, a platform built purely to track and improve how brands appear in AI answers, went from a $20 million Series A led by Kleiner Perkins to a reported $1 billion valuation in roughly a year, selling answer engine visibility to Fortune 100 marketers. G2 has rebuilt its research agenda around what it calls the answer economy. Even Forrester now puts "Ask AI about working with Forrester" buttons in its own website footer, pre-written prompts that send visitors to ChatGPT and Perplexity. When a research firm sends its own traffic to answer engines, the argument is over.
The Ignita insight: your launch is training data
Here is the conclusion most B2B marketers are missing. They are treating answer engine optimisation as a technical checklist: add schema, publish an llms.txt file, restructure the FAQ. Useful, but it optimises the 10% of the signal you control. The 90% is third party corroboration, and there is exactly one moment in a startup's life when it can manufacture a dense spike of independent coverage, reviews, analyst attention and social chatter all pointing at the same narrative: the launch.
A launch used to be a demand event. It is now a data event. The coverage you earn in launch week is scraped, indexed and cited by the models your buyers consult for years afterwards. A startup that launches quietly and plans to "do PR later" is choosing to be invisible to the 51% of buyers who start in a chatbot, because the model has nothing independent to cite. This is also why cultural brand building compounds now: every piece of named, third party evidence that your brand stands for something specific becomes machine readable reputation. The AI vendor shortlist is not a ranking you win with keywords, it is a reputation you earn in public, and the launch is the cheapest place to earn it. That is the thinking behind how we build launch narratives and brand programmes at Ignita.
What founders and marketers should do
Audit your AI presence this week. Ask ChatGPT, Claude, Perplexity and Gemini the questions your buyers ask: best tools in your category, alternatives to the category leader, your brand name directly. Log what is cited, what is wrong, and who beats you.
Treat your launch as a corroboration event. Concentrate press, reviews, analyst briefings, founder content and community activity into a tight window so the models ingest one coherent narrative. Study how the strongest raises convert funding news into durable coverage in our launch teardowns and insights.
Earn third party mentions before polishing owned content. One named comparison article, G2 review cluster or journalist writeup does more for answer engine visibility than another gated whitepaper.
Write answer shaped content. Question form headings, answer in the first sentence, specific numbers with sources. Models cite pages that resolve queries cleanly.
Arm humans for the validation step. Gartner's 69% will bring AI output to your sales team. Train reps to correct the record gracefully and publish the corrections publicly so the models learn them too.
Ship a real trial. With more than 60% of buyers using trials to de-risk decisions, a self serve trial is now part of your trust architecture, not just your funnel.
FAQ
What share of B2B buyers use AI to research vendors? 94% of business buyers use AI somewhere in the buying process according to Forrester's Buyers' Journey Survey, 2025. G2 found 71% of software buyers use an AI chatbot during research and 51% start there. Gartner's narrower measure found 45% used generative AI in a specific recent purchase.
Do younger buyers actually decide B2B purchases? Yes. Forrester found millennials and Gen Z passed 64% of business purchase influencers in 2022 and 71% in 2023, and the share only grows each year as older cohorts retire from buying committees.
Do B2B buyers trust AI recommendations? Partially. G2 found 83% feel more confident deciding with AI help, but Gartner found 51% expect misleading information from generative AI and 69% validate AI insights with a sales rep before acting.
What is answer engine optimisation? Answer engine optimisation, or AEO, is the practice of improving how a brand is represented and cited in AI generated answers from tools like ChatGPT, Perplexity and Copilot. It shifts effort from ranking pages to being accurately cited by models.
Is SEO dead for B2B startups? No, but its job changed. Search still feeds the models and captures late stage intent. The early research phase, where shortlists form, has moved into conversational AI, so visibility there now leads the priority order.
How does a launch affect AI visibility? Answer engines lean on independent sources: press, reviews, analyst notes and community discussion. A launch generates the densest spike of such coverage a startup ever gets, so launch week coverage effectively becomes the model's long term memory of your brand.
Keep going. If you want one sharp, data led read like this each week on how startups launch and win attention, join Ignita's free Substack at ignitaai.substack.com.
