Your Buyers Now Ask a Chatbot Before They Ask You. G2 Just Measured It.

What the 2025 Buyer Behavior Report says about AI search, shrinking shortlists, and how to be the answer instead of the afterthought.

---

Law Smith

Founder & President, Tocobaga

Law Smith is the Founder and President of Tocobaga, a Tampa-based ROI-focused marketing agency and SMB advisory. He has strategically advised 1,000+ small and medium businesses, executed 600+ integrated marketing campaigns, and written 30+ business plans. Tocobaga is a Google Partner and Squarespace Gold Partner headquartered in Ybor City.

In 1895, if you lived on a farm in Kansas, you did not shop. You asked the book.

The Sears catalog, 532 pages that spring, eventually thousands, was not a store. It was an answer engine made of paper. You had a question: what stove, what plow, what wedding dress, what price. The book had already done the comparing, the vetting, the ranking. Rural America outsourced its buying research to an intermediary it never met, and Sears owned American commerce for the better part of a century because it owned the moment the question got asked.

Then the front door moved. Interstates, malls, and eventually a website named after a river. Sears, the company best positioned on Earth to become the internet’s answer engine, with the catalog data, the logistics, the trust, watched the question move somewhere else and kept printing the book. It shipped its last big catalog in 1993, the year before Bezos incorporated.

Which has nothing to do with your 2026 pipeline. Except that the front door just moved again, and this time somebody actually measured the foot traffic.

The short version: G2 surveyed 1,169 B2B decision makers in April 2025. Generative AI chatbots are now the single most-cited external influence on vendor shortlists (17.1%, roughly double salespeople and Reddit-style forums) and the most-trusted source at decision time in North America (17.2%).1 Per G2’s own platform data, leads that arrive from AI search convert about 40% better than leads from traditional search.1 If an LLM cannot find, understand, and recommend your business, you are not losing deals to a competitor. You are absent from the room where the shortlist gets made. The fix is generative engine optimization: open your site to AI crawlers, restructure content into answer-first chunks with clean schema, and build the review and citation footprint the models actually read.

What Did the G2 2025 Buyer Behavior Report Actually Find?

Direct answer: that AI stopped being an experiment and became a standing line item, in budgets, in the buying journey, and in what buyers demand from the products themselves. G2’s framing for the year is that AI now stands for "Always Included."1

The spend picture first. 57% of buyers expect their organization to spend more on tech and software over the next year, up eight points from the prior report, while the share expecting to spend less nearly halved, from 9% to 5%.1 More telling than the amount is the funding source: more than half of AI software purchases are now funded from central IT budgets rather than R&D or discretionary experiment money, nearly 70% at enterprises with 1,000 to 5,000 employees.1 Experiments live on discretionary budgets. Operations live on IT budgets. AI moved buildings.

The value picture backs it up. Four out of five buyers reported positive returns on their AI-powered software investments, with 32.5% saying results exceeded expectations against only 7.8% falling short.1 Two out of three would pay a premium for AI functionality if the value is demonstrated, and among self-described power users, that jumps to 88%.1 Nearly half of enterprise buyers (49.5%) switched a software vendor in the past year specifically to get better AI features.1 Buyers are not humoring this stuff. They are re-shopping their whole stack over it.

Where Do Buyers Research Now, Google or AI?

Direct answer: both, but the mix is tilting fast. 79% of surveyed buyers say AI search has changed how they conduct research, and 29% already start their research with AI search more often than with Google.1 At enterprise and large-enterprise companies, software review sites and AI search have leapfrogged Google as the top research sources, review sites hit 61% and AI search 57% among 1,000–5,000-employee firms, against 53% for Google.1

Downstream, the shortlist is where it gets expensive to be invisible. Buyers now prefer narrowing to just two or three vendors, and the no-shortlist-at-all approach, straight from research to pick, has doubled in recent years.1 And the sources shaping that shortlist? GenAI chatbots lead at 17.1%, ahead of software review sites at 15.1%, vendor websites at 12.8%, and roughly double the influence of vendor salespeople (8.8%) and independent forums (7.6%).1

This rhymes with what Gartner is seeing from the other side of the table: 67% of B2B buyers now prefer a rep-free experience for at least part of the purchase, up from 61% a year earlier, and 45% used AI during a recent buying decision.2 The research is happening. It is just happening without you in the room, unless the machine doing the answering knows who you are.

If you read our Florida law firm playbook, this is the same story one octave lower. There, 78% of legal searches ended in an AI answer before anyone clicked. Here, the whole B2B economy is running the same play. The searcher changed species: part human, part model. You now market to both.

Who Actually Makes the Buying Decision in 2026?

Direct answer: fewer people, faster, with a department lead holding the pen. Buying committees are shrinking, trios and quartets grew 9% year over year while the classic five-to-eight-person committee fell 11%.1 Department leads are now the single most common final decision maker at 24%, ahead of IT at 22% and the CFO at 20%.1 IT sits at the table in nearly half of all purchases (47%), and roughly three in four companies apply stricter security and legal review to AI-powered software than to regular software.1

Two implications for anyone selling anything. First, the internal champion is now the whole ballgame, small teams move on the recommendation of one convinced person, so your content has to arm that person to win the internal argument. Second, buyers want salespeople later: 62% now prefer rep contact at the evaluation and decision stages rather than up front, a 17-point jump in one year.1 By the time they talk to you, the model and the reviews already briefed them. Your job is to have written the briefing.

What Does "AI Always Included" Mean for Deals and Pricing?

Deal sizes are settling smaller. Multi-year mega-deals are shrinking, enterprises now cluster purchases around $100K–$150K, and mid-market buyers have adopted a do-more-spend-less-per-deal posture.1 More than one in three buyers now prefer variable pricing for AI-era software, pay-as-you-go leads emerging models at 39%, with usage-based at 28% and outcome-based at 10%.1 Land small, prove value, expand. Which, not for nothing, is exactly the argument against signing a twelve-month auto-renewing agency retainer before anyone has defined a qualified lead. Buyers get it. Vendors are adapting. Agencies should too.

How Does a Business Show Up in AI Search? (The GEO Checklist)

Direct answer: generative engine optimization, GEO, is SEO’s successor discipline for a world where the machine gives an answer instead of a list of links. G2’s own recommendations land on the same playbook we run:1

1.  Unblock the crawlers. Your site may be invisible to AI on purpose and by accident: a blanket Disallow in robots.txt, a GPTBot or Google-Extended block, "noai" or "noindex" meta tags flipped on by a CMS template, canonicals pointing at stale URLs, or key content rendered in JavaScript the crawler never executes. Fifteen minutes in your robots.txt and page templates settles whether you are even eligible to be an answer.

2.  Restructure content for answers, not rankings. LLMs lift chunks: direct answers under question-shaped headings, clean FAQ schema, plain-English definitions, quotable stat lines with sources. Convert product docs, service pages, and FAQs into answer-first blocks. This is the exact format this post is written in, the format is the strategy.

3.  Build the citation footprint. Recent peer reviews and brand mentions on category and comparison pages carry real weight in what LLMs retrieve and recommend, G2’s data shows review platforms outperform forums and analyst firms on LLM mentions. For local and service businesses, the same logic runs through Google Business Profile, industry directories, and earned press.

4.  Monitor your presence in the models. Ask the major chatbots the questions your buyers ask, "best [category] for [industry] in [city]", monthly. Log who gets named, who gets cited, and what the models get wrong about you. You cannot fix an answer you have never read.

One receipt from our own scoreboard: when we pulled our Search Console AI-visibility data this month, our online reputation management page was our second-most-cited page in generative AI results, because it is structured exactly this way. We run this playbook on ourselves before we sell it to anyone.

What Should an SMB Do This Quarter?

1.  Audit AI-crawler access this week (robots.txt, meta robots, CDN bot protection). Cost: an hour. Risk of skipping: total invisibility.

2.  Rewrite your top three money pages answer-first: a direct answer in the first two sentences under every question-shaped heading, FAQ schema in JSON-LD, no duplicate inline microdata.

3.  Get ten fresh, real reviews on the platform your buyers trust, G2 or Capterra for software, Google for local service, industry directories for professional services.

4.  Run the chatbot test monthly and keep a log. Treat "what the models say about us" as a KPI next to rankings and traffic.

5.  Arm your champion. Build the one-pager your internal advocate forwards: outcomes, proof, pricing logic, security answers. Small committees decide fast, be easy to argue for.

Tocobaga is a Tampa-based, ROI-focused marketing agency and SMB advisory. We have advised 1,000+ SMBs and executed 600+ campaigns, and AI-search visibility is now baked into every SEO and content engagement we run, same flat fee, no percentage of ad spend, no auto-renew, and your content stays yours. If you want to know whether the machines can even see you, the AI Visibility Audit is free. If we are not the right fit, we will point you to someone who is. Book an intro call at SolvingHow.com or call (813) 934-6605.

What is generative engine optimization (GEO)?

GEO is the practice of making your business findable, understandable, and recommendable by AI systems: ChatGPT, Gemini, Claude, Perplexity, and AI Overviews, rather than just rankable by traditional search engines. Search engines return links; answer engines return answers. GEO covers crawler access, answer-first content structure, schema, and the review and citation footprint the models draw on.

Do AI-search leads actually convert?

Per G2 platform data in the 2025 Buyer Behavior Report, leads generated through AI search convert roughly 40% better than leads from traditional search, largely because buyers complete much of their product-fit analysis before they ever reach your site.

Should I block AI crawlers to protect my content?

For most SMBs, blocking AI crawlers on marketing pages is self-sabotage: it removes you from the answers your buyers read while doing nothing to stop competitors from filling the space. Gate genuinely proprietary material; open everything designed to win customers.

Is this only a software-industry problem?

G2 surveyed B2B software buyers specifically, but the behavior generalizes. Our own Search Console data shows generative-AI impressions on service-business pages, and legal-industry research shows the same pattern in consumer searches. Assume your buyers ask a chatbot; verify by asking it yourself.

What is the fastest GEO win?

Crawler access plus one answer-first rewrite of your highest-intent page, with FAQPage schema. It is unglamorous, it takes a day, and it determines whether you are eligible to be cited at all.

Sources

1. G2, '2025 Buyer Behavior Report: AI Now Means Always Included,' Tim Sanders (author), 2025. Survey of 1,169 B2B decision makers across North America, EMEA, and APAC, conducted April 2025; supplemented with G2 platform data. https://research.g2.com/
2. Gartner, 'Gartner Sales Survey Finds 67% of B2B Buyers Prefer a Rep-Free Experience,' press release, March 9, 2026 (survey of 646 B2B buyers, August-September 2025). https://www.gartner.com/en/newsroom/press-releases/2026-03-09-gartner-sales-survey-finds-67-percent-of-b2b-buyers-prefer-a-rep-free-experience