AI now calls the shots in B2B buying
AI-powered search has moved from side project to standard practice in UK B2B buying. In a study of 175 UK decision-makers (Norstat for Clarity Global), 79% use AI weekly or daily at work, and one third use it every day. 64% spend one to four hours a week using AI to make decisions; 80% spend at least an hour. AI isn't a research aid anymore - it's part of the workflow that decides what gets read, shortlists vendors, and builds the internal case.
Discovery has compressed. Between 52% and 59% of buyers rely more on AI summaries, use traditional search less, visit fewer sites, and read fewer long articles. 59% say they now spend less time gathering raw information and more time judging what an LLM gives them. Your brand meets buyers through a model's summary, not your homepage.
The AI chooses your brand for them
At the top of the funnel, 87% of buyers use AI outlines to decide what to read. During shortlisting, 65% lean on AI for vendor selection. In evaluation, 77% substitute AI for due diligence and technical assessment. Inside the org, 75% use AI to create or influence the business case.
The implication is blunt: influence depends on how models interpret and summarise your brand. Volume matters less than clarity, structure, and evidence that software can parse without confusion.
GEO over volume: work for the summary
Generative Engine Optimisation (GEO) is newer and less predictable than SEO. AI search is a black box - updates, training data, and answer logic are opaque. Treat tactics as hypotheses, track outcomes, and adjust. No silver bullets.
- Write for summarisation: Clear claims, short sections, FAQs, and direct answers to common prompts. Make your "why us" scannable.
- Make claims verifiable: Use specific numbers, cite independent sources, and avoid vague superlatives.
- Keep experts visible: Include named subject-matter experts, quotes, and data-backed POVs.
- Strengthen technical signals: Fast pages, clean IA, and structured data so machines can extract meaning. See Google's guidance on structured data here.
- Standardise language: Use consistent terminology across web pages, PR, sales decks, and spokesperson comments.
Channel mix: where AI gets its evidence
Models pull from a messy blend of owned content, search visibility, social signals, and third-party validation. Early-stage answers lean more on third-party sources than your site. Late-stage prompts like "best" or "top" give weight to information that's corroborated on the open web, with third-party coverage treated as less subjective than vendor claims.
- Prioritise earned authority: Analyst notes, reputable media placements, and topical newsletters. Quality over link volume; fresh citations matter.
- Build social proof: Independent reviews (e.g., G2, Capterra), detailed case studies, and customer quotes that include concrete outcomes.
- Create comparison-ready assets: Transparent pricing explainers, implementation guides, and head-to-head pages that models can summarise cleanly.
- Maintain message integrity: The same claims and terms across every channel reduce ambiguity for AI and buyers.
Spend reallocation: where the ROI likely is
- Keep technical SEO healthy: Performance, structured data, crawlability. Models still need clean inputs.
- Elevate content with proof: Named experts, third-party data, original insights, and impartial reviews.
- Lean into PR and analyst relations: Target relevance and authority, not just links. Aim for citations that get quoted often.
- Invest in review programs: Encourage authentic, detailed customer feedback and keep it current.
Measurement that admits uncertainty
Owned analytics can't show how you appear in AI answers. Those responses are ephemeral - they vary by prompt, model, time, and context. The mandate: monitor, log, and iterate.
- Track AI answer surfaces for priority queries weekly. Note whether models reproduce your key claims or just mention the brand.
- Define a brand card: 5-7 canonical claims with proof points. Check for consistency across AI answers and refine content accordingly.
- Use automation to simplify AI search monitoring with real-time metrics (the study cites the need, not a specific tool).
- Run a quarterly review loop: Clarify messaging, update assets, refresh third-party citations, and re-measure.
What this means for marketing leaders
- Treat AI search as a primary discovery channel.
- Align messaging and proof so LLMs can extract, verify, and repeat your claims without distortion.
- Shift spend toward integrated content, third-party validation, and technical fundamentals that models can trust.
The sample is limited (175 UK decision-makers), but the signal is clear: AI is now part of the B2B buying process from first click to final justification.
Next step: upskill your team
If your team needs practical training on GEO, AI search, and prompt-to-pipeline execution, explore the AI Certification for Marketing Specialists or browse the latest AI courses. Managers responsible for adoption can follow the AI Learning Path for Business Unit Managers to align strategy, governance, and operational adoption across teams.
Your membership also unlocks:
