The short answer: yes, and arguably more than before, but not for the reason review platforms like to give. In TrustRadius's 2026 survey of 1,862 technology buyers, 63% used AI during their purchase journey and 94% of those fact-checked its answers at least some of the time. 74% used reviews to inform the decision. In G2's 2026 buyer research, review sites (38%) and AI chatbots (37%) were the two most-cited influences on the shortlist. AI is how buyers build a shortlist faster. Reviews are how they check it. Both datasets come from review platforms, so read them with that in mind.
If your pipeline increasingly starts with "we asked ChatGPT for options", it is fair to wonder whether the review profile you neglected for two years still matters. Two large 2026 surveys say it does.
In this post, a review site means a third-party platform where verified customers rate business software, such as G2, TrustRadius or Capterra. It is different from testimonials on your own website, which you control.
What does the 2026 data actually say?
Two surveys, both fielded in 2026, both large enough to take seriously:
| TrustRadius 2026 B2B Buying Disconnect | G2 2026 buyer research | |
|---|---|---|
| Who | 1,862 technology buyers and 444 vendors, global | 1,076 software buyers (March 2026, "Answer Economy"); more than 1,000 buyers plus 50+ interviews (2026 Buyer Behavior Report) |
| Published | 15 July 2026 | 15 April 2026; July 2026 |
| AI in the journey | 63% used AI during their purchase journey | 51% start research with an AI chatbot more often than Google, up from 29%; more than 80% got recommendations from an AI chatbot in the past two years |
| Trust in AI | 94% of AI users fact-check it at least some of the time | 61% use AI search alongside Google |
| Reviews | 74% use reviews to inform the decision | Review sites (38%) and AI chatbots (37%) top the shortlist influences |
| Other signal | Analyst reports used by 13%, down 63% since 2022; 83% shortlisted three or fewer products | IT security review is the biggest delay after selection (39%) |
Read the rows together and one sequence appears. AI narrows the field fast. Then buyers turn to sources they trust to check the answer: demos, trials, people they know, and reviews.
Why should you read this data sceptically?
Because TrustRadius and G2 are review platforms, and both surveys found that reviews matter. TrustRadius has been owned by HG Insights since June 2025; G2 sells review programmes to software vendors. That doesn't make the numbers wrong. It does mean you should separate what buyers did from what the platforms conclude.
- Behavioural numbers are the strongest. "63% used AI", "83% shortlisted three or fewer", "13% used analyst reports". These are answers about what a buyer did on a recent purchase.
- Interpretive claims are the weakest. TrustRadius's release says AI recommendations are "heavily influenced by trusted third-party content — including customer reviews". That may well be true, but it is the platform's reading, not a measurement in the survey.
- Different surveys measure different things. Forrester's 2026 study found 94% of buyers used AI in their most recent purchase, as covered in how B2B buyers now use AI to choose vendors. TrustRadius found 63%. Neither is wrong. They asked different people different questions. Quote the survey whose question matches your point.
How do reviews and AI answers connect?
Think of a B2B purchase in 2026 as two filters.
Filter one: the shortlist. A buyer asks an AI assistant for options in a category. The assistant writes its answer from whatever it can find: your site, comparison articles, forums, and review platforms that are open to crawlers. If you're thin everywhere outside your own site, you're thin in the answer. With 83% of TrustRadius's buyers shortlisting three or fewer products, missing the shortlist means missing the deal.
Filter two: the check. The 94% fact-check figure is the important one. A buyer who got your name from ChatGPT goes looking for confirmation. Recent, specific, verified reviews from companies like theirs are the cheapest confirmation you can give. Two-year-old five-star ratings with no text are not.
Analyst reports are the counter-example. At 13% usage, down 63% since 2022 in TrustRadius's data, the expensive third-party validation has lost ground to the cheap, plentiful kind.
What should a $3M–$20M B2B company actually do?
Don't buy every platform's premium profile. Do the work that helps both filters:
- Pick one review platform, not three. Ask your last ten closed-won customers where they researched. Concentrate reviews there. A thin presence across three platforms helps nobody.
- Ask for reviews at the right moment. After onboarding succeeds or a renewal is signed, not at the sales handover. Ask the champion, the person who will be asked "how is it going?" by peers.
- Make reviews specific. Ask reviewers to say what they replaced, their company size and what changed. "Easy to use" is noise. "Replaced spreadsheets for a 40-person ops team" is evidence.
- Answer the comparison question on your own site. AI assistants compare you with competitors whether you like it or not. A plain page saying who you are and aren't right for gives them something accurate to use. The same principle drives getting named in AI answers for local businesses.
- Publish your pricing, or at least a range. Transparent pricing has topped TrustRadius buyers' wish list for four years running. A buyer checking an AI shortlist will drop the vendor whose price they can't find.
- Prepare for the security review. G2 found IT security review is the biggest delay after a vendor is chosen. A ready security page or trust centre shortens the part of the deal reviews can't help with.
Common questions
Do B2B buyers trust AI recommendations for software?
They use them but check them. In TrustRadius's 2026 survey of 1,862 technology buyers, 63% used AI during their purchase journey, and 94% of those said they fact-check its answers at least some of the time. AI builds the shortlist; demos, trials and reviews confirm it.
Are G2 and TrustRadius reviews still worth investing in?
For most B2B software companies, yes, within limits. G2's 2026 research found review sites were the top influence on shortlist decisions (38%), just ahead of AI chatbots (37%). Focus on one platform your customers actually use, and on recent, specific reviews rather than volume.
Why do surveys disagree on how many B2B buyers use AI?
They ask different people different questions. Forrester's 2026 study found 94% of buyers used AI in their most recent purchase; TrustRadius found 63% used AI during their purchase journey. Sample, wording and definition of "using AI" all differ. Quote the figure whose question matches your point.
Do analyst reports still matter in B2B buying?
Much less than they did. TrustRadius's 2026 survey found only 13% of buyers used analyst reports to make purchase decisions, a 63% fall since 2022. Demos, free trials, prior experience and user reviews ranked among the most influential resources instead.
How many vendors do B2B buyers shortlist?
Few. In TrustRadius's 2026 survey, 83% of buyers shortlisted three or fewer products. Combined with AI-assisted research, that means the shortlist often forms before a vendor knows a buyer exists. Being absent from that early list is close to being absent from the deal.
The honest summary
AI didn't make reviews obsolete. It moved them to a different job. They're less how buyers discover you and more how buyers check what an AI told them. The surveys that say so come from companies that sell review programmes, so keep the behavioural numbers and discount the conclusions. The practical response is modest: one platform, recent and specific reviews, honest comparison and pricing pages, and a security page ready before the deal stalls.
Where reviews fit in your wider demand system — alongside content, paid and sales — is the sort of thing a marketing systems consultant maps out. If your budget is still being set, start with how much a B2B company should spend on marketing.
Sources: TrustRadius 2026 B2B Buying Disconnect Report release (15 July 2026) · G2 — The Answer Economy (15 April 2026) · G2 2026 Buyer Behavior Report
Written by Jagatjeet, a web architect and digital marketer based in Kamloops, BC. Jagatjeet (jagatjeet.com) advises B2B companies on marketing systems and web architecture. Published 12 October 2026.