Forrester's latest forecast says AI agents in B2B buying will play a part in three out of four business purchases in 2027.1 For heads of demand generation and sales leaders, that raises an awkward question: if a buyer's first shortlist is assembled by software, how does your company get on it, and what happens to the leads, meetings and pipeline you depend on? The research published over the past few months gives a clearer answer than the headlines suggest. AI is changing how buyers find and filter suppliers, but it is also making them more sceptical, and that scepticism sends them back to people.
What Forrester predicts for AI agents in B2B buying
On 21 October 2026 Forrester published its 2027 predictions for B2B go-to-market teams.2 The headline forecast is that "agentic buying" (buyers delegating parts of the purchase to AI agents) will play a role in 75% of B2B purchases in 2027.1 Forrester expects agents to be used from supplier discovery through to analysing contract renewals, but it also says fully autonomous buying will stay concentrated in high-volume, low-risk, repeatable purchases.2 Complex, expensive decisions will still involve people.
Three of the predictions matter most for pipeline:
- Buyers will double their human interactions to check what AI tools recommend, turning to supplier teams and trusted third parties such as analysts, customer references and partners.2
- Two-thirds of providers will lose deals because buyers cannot verify their claims. Forrester's analysts argue that claims need proof points that AI tools can find and check.3
- Many CMOs will overestimate AI productivity. Forrester expects some marketing leaders to cut as many as 25% of roles in anticipation of gains that do not arrive, then face a choice between cutting programmes and running them with less capacity.2
These are predictions, not measurements. Forrester's public pages do not publish a method behind the 75% figure, so treat it as an informed forecast of direction rather than a precise number.
B2B buyers already research with AI, then check with people
The forecast builds on behaviour that is already visible in survey data. Forrester reported in January that 94% of business buyers say they use AI somewhere in their buying process, up from 89% a year earlier, and that 61% use private AI tools provided by their own organisation.4 That second number matters: much buyer research now happens inside a company's own Copilot or ChatGPT environment, where your website analytics and intent signals cannot see it.
Gartner's survey of 645 B2B buyers, run in August and September 2025 and published in May 2026, adds the other half of the picture.5 Buyers used an average of seven information sources during a recent purchase, and 45% had used generative AI, mainly to gather information on vendors and products. Yet trust is thin: 51% said they were more likely to meet misleading information from GenAI, and almost as many (49%) said the same about sales reps. The result is a habit of cross-checking, with 69% preferring to validate AI-generated insights with a sales rep.
Vendor research points the same way for software purchases. G2, which runs a software review platform, surveyed 1,076 B2B software buyers in March 2026.6 Half (51%) said they now begin their purchasing process in an AI chatbot rather than a search engine, and 69% said they ended up choosing a different vendor from the one they first planned because of chatbot guidance. About a third bought from a vendor they had never heard of before.
Taken together, AI is reshuffling shortlists before a seller is involved, and then buyers want a human to confirm the reshuffle was sound. If you have read our earlier piece on what B2B buyers are telling us in 2026, this is the next step in the same trend: the shortlist forms even earlier, and it is shaped by machines as well as people.
Proof beats claims when AI does the first check
Forrester's warning that two-thirds of providers will lose deals over unverifiable claims is the most practical of its predictions.3 When a buyer asks an AI assistant to compare three suppliers, the tool summarises what it can find and check. Vague superlatives, gated assets with nothing indexable, and messaging that differs between your website, your sales deck and your outbound emails all weaken your case.
This has a direct link to demand generation. Content syndication, webinars and analyst-style research are not just lead capture mechanisms any more; they are also the material AI tools draw on when buyers ask questions. Content that names specific use cases, explains how a product works and shows evidence a buyer can check is more useful to a human reader and to an AI summary. It also produces better leads, which we cover in how to judge content syndication lead quality.
Agentic buying still needs human touchpoints
A common reaction to agentic buying forecasts is to assume outbound calling, SDR teams and appointment setting are about to become obsolete. The evidence does not support that. Forrester expects human interactions to rise, not fall, because buyers want someone to vouch for what their AI tools told them.2 Gartner's data shows the same preference for validation from a sales rep.5
What changes is the job those touchpoints do. A first call or email is less often the buyer's introduction to your category, and more often a check on something they have already read. That favours reps who can answer specific questions accurately, qualify properly, and pass a well-documented conversation to an account executive. It penalises scripted volume, inaccurate claims and outreach that ignores what the buyer already knows. A well-run SDR function or appointment-setting programme becomes part of how buyers verify you, not just how you find them.
There is also an operational warning in the CMO prediction. If teams cut people on the assumption that AI will fill the gap, the human capacity buyers are asking for may not be there when they want it.2
What this means for your pipeline
- Audit your claims for verifiability. List the ten claims your sales team makes most often. For each, check whether a buyer (or their AI assistant) could find public evidence for it. Rewrite or remove anything that cannot be supported.
- Make your messaging consistent. Compare your website, sales deck, outbound scripts and syndicated content. Where AI tools find conflicting statements, buyers find doubt.
- Publish content that answers buying-group questions. Specific, ungated explainers on pricing models, implementation and integration give AI tools accurate material to summarise. Keep gated assets for deeper research and use content syndication to reach the right job functions with them.
- Train callers and SDRs as validators. Expect prospects to arrive with AI-generated comparisons. Give reps accurate answers to common objections and permission to say "I'll check and come back", rather than guess.
- Tighten qualification. When buyers come with partial research, qualifying questions on authority, need and timing matter more. Record what the prospect already believes about you so the next conversation starts from there.
- Protect human capacity before cutting it. Model what happens to meeting volume and follow-up speed if headcount falls before AI productivity is proven in your own data.
Frequently asked questions
What is agentic buying in B2B?
Agentic buying means a business buyer delegates parts of a purchase to AI agents, such as finding suppliers, comparing options, summarising reviews or analysing contract renewals. Forrester predicts it will play a role in 75% of B2B purchases in 2027, but expects fully autonomous buying to stay limited to low-risk, repeatable purchases. Complex deals will still involve people.
Will AI agents replace SDRs and outbound sales teams?
Current evidence says no. Forrester predicts buyers will double their human interactions to validate AI recommendations, and Gartner found 69% of buyers prefer to check AI-generated insights with a sales rep. The role is shifting from introducing a category to answering specific questions accurately, qualifying well and passing informed conversations to account executives.
How do I make sure AI tools recommend my company to B2B buyers?
No one can guarantee AI recommendations. What helps is clear, consistent and checkable information: specific product and use-case content, claims backed by public evidence, and the same message across your website, sales materials and third-party sources. Forrester warns that providers whose claims buyers cannot verify will lose deals.
A note on the evidence
Forrester's 2027 figures are forecasts, and its public summaries do not give the methodology behind them. Its 94% AI-usage figure comes from a blog post that does not state sample size or fieldwork dates. Gartner's survey has a stated sample of 645 buyers, but its fieldwork took place in 2025 and the release gives limited respondent detail. The G2 figures are vendor research from a software review platform, cover software buyers only, and are reported here via Demand Gen Report. The two charts use different surveys and should not be compared directly.