TopicSEO
B2B SEO in 2026 works in two layers. An authority layer of citable content wins mentions in AI Overviews, AI Mode and ChatGPT. A conversion layer of service, comparison and pricing pages captures the high-intent clicks AI cannot satisfy. Judge the programme on pipeline, not traffic. This guide covers the full system.
What changed, and the two-layer strategy that follows from it
The numbers first, because they explain every strategic choice in this guide. Google AI Overviews now appear on roughly 48% of queries by Semrush's count, and Seer Interactive's dataset has click-through on affected queries down by around 61%. AI Mode serves over 75 million daily users, according to Alphabet's Q3 2025 results, and Semrush's session study found that 92 to 94% of its sessions end without a click. ChatGPT, at around 900 million weekly users per OpenAI's February 2026 disclosure, is now where a great deal of B2B research begins. And yet even SparkToro's 2026 zero-click study, the bleakest count going, found that 68% of searches end without any click, which leaves roughly a third still sending one. Organic remains the largest controllable pipeline source for most B2B firms. SEO is not dead. The informational click is the part that is dying.
The decline is not spread evenly. AI answers absorb the top and middle of the funnel. Definitions, comparisons of approaches, "how does X work", early research. What they cannot absorb is the bottom. A buyer who needs to see your pricing, compare you against a named competitor and fill in a form still has to visit your site, and those queries retain strong click-through. Meanwhile, buyers now shortlist vendors straight from the AI answer itself. Referrals from AI answers grew +527% year on year, comparing January to May 2025 with the same months in 2024, in the dataset Search Engine Land reported. Semrush's analysis values them at about 4.4 times an organic visitor by conversion rate, because they arrive pre-recommended.
Hence the two-layer strategy that organises the rest of this guide. The authority layer exists to be cited. Statistics pages, pricing transparency, honest comparisons, research, all engineered so AI engines quote and recommend you during the buyer's invisible research phase. The conversion layer exists to be clicked. Service pages, competitor comparisons, case studies and pricing pages that capture the high-intent traffic AI hands off. Most B2B sites have a weak version of the second layer and nothing at all in the first. Building both, and linking them to each other, is what B2B SEO now means.
Keyword research that maps to revenue, the intent tiers
B2B keyword research goes wrong by chasing volume. The queries with impressive numbers are almost always informational, exactly the ones AI now answers without a click. The queries that close deals look pathetic in a keyword tool. "n8n consultant manchester", 30 searches a month. The correction is to organise research by intent tier and work bottom-up, starting with the tier that pays for itself and using it to fund the rest.
Tier 1, buying intent. "[service] agency", "[service] consultant [city]", "[competitor] alternative", "best [category] for [industry]". Tiny volumes, direct pipeline. These map one-to-one to conversion-layer pages and get built first. A single ranking here can pay for the whole programme. Tier 2, evaluation intent. "[service] cost", "[tool A] vs [tool B]", "[approach] pros and cons", "how to choose [category]". Moderate volume, high citation potential. These queries trigger AI answers, which makes them the natural home of the authority layer. The page cited in the answer wins the shortlist. Tier 3, problem intent. "why is [metric] falling", "how to [outcome]". Higher volume, mostly zero-click now. Build selectively, only where a piece includes a statistic or framework strong enough to be cited and to earn links.
How strong is your own AI marketing? Take the free two minute scorecard and get an honest read plus the next steps that matter most.
Take the scorecardTwo research moves specific to B2B repay the effort. First, mine your own pipeline. The phrases prospects use on discovery calls, in tender documents and in "how did you find us" answers are tier-1 and tier-2 queries no tool surfaces. Your CRM is a keyword database. Second, map the committee rather than just the searcher. The technical evaluator googles implementation questions, the finance approver googles cost and ROI, the end user googles ease of use. A deal requires content answering all three, which is why thin sites lose late-stage deals they never knew they were in.
| Tier | Example queries | Volume | Role | Layer |
|---|---|---|---|---|
| 1 | Buying | "[service] agency", "[competitor] alternative" | Low | Direct pipeline | Conversion |
| 2 | Evaluation | "[service] cost", "[A] vs [B]" | Medium | Shortlisting + AI citation | Authority |
| 3 | Problem | "how to improve [outcome]" | High | Selective. Citation and links only | Authority |
Technical foundations, the unglamorous entry ticket
Technical SEO decides whether anything else in this guide gets the chance to work, and in 2026 it serves two sets of crawlers at once. The classic checklist still applies. Clean crawlability and indexation (no orphaned pages, no parameter sprawl, intentional canonicals), site speed within Core Web Vitals thresholds, HTTPS, mobile rendering, and an XML sitemap that reflects reality. For most B2B sites, typically hundreds of pages rather than millions, this is something you fix once and then audit every quarter. It should not absorb the budget that content and authority need.
The AI-era additions are small but consequential. Audit robots.txt for the AI crawlers (GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended), because a 2023-era block quietly removes you from the answer layer entirely. Verify the site in Bing Webmaster Tools, because ChatGPT retrieves primarily through Bing's index. Ensure important content renders without JavaScript execution, because several AI crawlers read raw HTML. And keep schema honest and complete. Organization with sameAs, Service, Person for authors, FAQPage where genuine Q&A exists.
One mistake B2B sites make again and again is gating content without asking who will be able to read it. Put a substantial asset behind a form and no crawler, search or AI, can reach it. The whitepaper PDF nobody can crawl earns no authority, no citations and no AI mentions. The 2026 pattern is ungated HTML for the material you want engines and buyers to find (statistics, methodology, comparisons). Reserve gating for custom artefacts like calculators and templates, where the trade is obvious to the visitor.
- Classic hygiene, crawlability, indexation, Core Web Vitals, HTTPS, honest sitemap.
- AI crawler access, GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended unblocked.
- Bing Webmaster Tools verified, the ChatGPT retrieval path runs through Bing.
- Important content renders without JavaScript. Schema mirrors visible content.
- Citable assets published as ungated HTML, not locked PDFs.
The authority layer, content formats that get cited
The authority layer has one job. When an AI engine, a journalist or a prospect's analyst needs a fact about your category, your page is the one that gets quoted. Citation studies are unambiguous about what earns this. Adding statistics, quotations and citations boosted visibility in generative answers by up to 40% in the GEO research paper that named the field, and Seer's 2026 recency study found 75% of the pages LLMs cite were updated within the last year. Authority content is engineered, not just written. Four formats do most of the work in B2B.
Pricing transparency pages, "how much does [service] cost in [market]", are the single most reliably cited B2B format, because cost is the question every buyer asks and almost every vendor dodges. Publishing real ranges with the variables that move them feels exposing. It is also how you become the source engines quote when anyone asks what the work costs, and that keeps you on shortlists. Statistics pages (curated, dated, refreshed industry numbers) become citation magnets for AI answers and human writers alike, and they compound for years. Honest comparisons ("[tool A] vs [tool B]", "best [category] agencies"), with methodology shown and your own bias declared, win tier-2 evaluation queries and get picked up by AI Mode at the same time. Original research, even a modest survey of your own customer base, produces numbers that exist nowhere else, which is the strongest reason anyone has to cite you.
Whatever the format, the engineering rules are constant. Open every section with a 40-60 word extractable answer under a question-shaped heading, include at least one table and one list, attach a named author and a visible updated date, and refresh on a calendar rather than on inspiration. The authority layer is also where GEO and SEO become one budget. Every property that makes a page citable by AI engines also makes it rank and makes it quotable by humans.
The conversion layer, pages that turn demand into pipeline
The conversion layer is where the surviving clicks land, and because AI-referred visitors convert so much better than ordinary organic traffic, every weakness here costs more than it used to. The core inventory for a B2B firm starts with one service page per offer per market, each written as a sales conversation rather than a brochure. Who it is for, what it costs (ranges beat silence), how engagement works, proof, and the questions a sceptical buyer actually asks, answered in FAQ blocks with schema.
Comparison and alternative pages attract the highest-intent organic traffic in B2B. "[competitor] alternative" and "[you] vs [competitor]" queries are typed by buyers actively assembling a shortlist. Write them honestly, acknowledge what the competitor does well, be specific about where you differ, and let the structure (a real comparison table, named criteria) do the persuading. Honesty is also self-interest. Dishonest comparison pages get named and shamed in the Reddit threads engines cite, which turns an asset into a liability. Case studies round out the layer. Make them specific, with real numbers, structured around problem to intervention to result, because "increased qualified pipeline 40% in two quarters" is both a conversion argument and an extractable claim.
Treat conversion-layer craft as seriously as acquisition. Clear next steps on every page (book a call, get a scoped quote, not "learn more"), forms that ask only what sales needs, response-time promises you keep, and "how did you hear about us?" capture so the measurement section below has data to work with. B2B buying committees revisit these pages repeatedly across a multi-month cycle. They are the most-read sales documents your firm produces.
- One service page per offer x market, priced in ranges, FAQ-schema'd.
- Comparison and alternative pages for every named competitor buyers shortlist.
- Case studies with numbers, structured problem to intervention to result.
- Pricing page or guide, the page buyers want and competitors withhold.
- Friction-free conversion, minimal forms, explicit next step, kept response promises.
Internal architecture, hub-and-spoke that compounds
Architecture is what makes the two layers work together rather than simply sit on the same site. The pattern that works is hub-and-spoke. A full hub page owns the head topic ("B2B SEO", "AI automation"), and spokes own every substantial sub-question, from cost and comparisons to how-tos and industry variants. Each spoke links up to the hub and across to relevant siblings, and the hub links down to all of them. Buyers experience a site that anticipates their next question. Crawlers experience a topic covered so thoroughly that authority consolidates instead of fragmenting.
The AI-era reason this matters more than it used to is query fan-out. Google's AI Mode decomposes a buyer's question into many sub-queries and retrieves for each. A firm whose cluster covers cost, alternatives, timelines and industry specifics appears in multiple retrievals for a single user question and dominates the synthesised answer. One monolithic page, however good, can win only the retrieval it exactly matches.
Three habits keep the architecture working in practice. Link with descriptive anchors ("UK GEO pricing guide", not "click here") so both engines and users know what is on the other end. Keep commercial pages within two or three clicks of every related authority page. The authority layer earns the visit. The internal link delivers it to the conversion layer. Prune as well. Consolidate overlapping posts that compete with each other, redirect the retired URLs, and let one strong page own each question. Most mature B2B sites improve more by deleting and merging than by publishing.
Digital PR and E-E-A-T, the trust layer
Roughly 96% of pages cited in AI Overviews show strong E-E-A-T signals (experience, expertise, authoritativeness, trust) in Wellows' analysis of 15,847 AI Overview results, and the same signals dominate classic ranking on commercial queries. For B2B firms that is easier than it sounds. You already employ genuine experts, and the work is making that expertise visible to machines. Named authors are the foundation. Every substantial page should be bylined to a real person with credentials, a bio page, Person schema and a linked LinkedIn profile. Anonymous content is a structural disadvantage in 2026. Engines cannot verify an expert who has no name, and "admin" has no E-E-A-T.
Digital PR has refocused from links to mentions. Unlinked brand mentions now correlate with AI citations more strongly than backlinks do, and Reddit and LinkedIn are the most-cited domains in AI answers. What other people say about you in public counts for more than what your own site says. Get your experts quoted in trade press (data and contrarian-but-defensible positions earn the invitations), publish on LinkedIn with substance rather than engagement bait, maintain an honest presence where your market actually discusses vendors while disclosing affiliation, and pursue the directory and "best of" listings buyers and engines both consult. Links still matter for classic ranking, but the mention, the review and the thread now carry just as much weight.
Experience, the first E, is worth showing deliberately, because it is the one signal a competitor cannot generate. Use first-person specifics ("across the forty audits you ran last year, the most common failure was..."), real client situations, named tools, actual numbers and photographs of actual work. Generic best-practice listicles are now free to produce, and worth very little.
Local and multi-market B2B
B2B firms underrate local search because their ambitions are national, but their buyers keep searching locally regardless. "[service] agency manchester" and "[category] consultant near me" remain stubbornly common, because buyers still prefer suppliers they could meet. The local work is cheap relative to what it returns. Complete the Google Business Profile and the Bing Places listing, since the ChatGPT pipeline runs through Bing. Fill in every category and service field, keep reviews coming in and reply to them, and write local pages that cover the office, the team and how to find it rather than running find-and-replace on a town name. With roughly 45% of consumers now using AI tools for local recommendations, per BrightLocal's Local Consumer Review Survey, the assistant answering "good [service] firm near me" relies directly on these signals.
Multi-market work, the UK firm selling into the US or the reverse, is mostly a matter of localisation. Give each market its own pages or section, with hreflang. Match the buyer's spelling, optimisation for the UK and optimization for the US. Quote prices in the local currency at figures that make sense there, because £1,500 to £5,000 a month reads differently from $2,000 to $6,500. Use proof, case studies and compliance vocabulary from that market, because UK GDPR and the ICO mean nothing to a Texan buyer, and SOC 2 means little in Sheffield. Engines resolve entities per market too, so consistent local profiles, local reviews and local mentions decide which firms get named when the prompt includes a place.
The sequencing rule for both is depth before breadth. One market with full two-layer coverage outperforms three markets of thin pages, and near-duplicate city pages are discounted by Google's quality systems in any case. Expand when the home market's cluster is built, cited and converting.
Measurement means pipeline, not traffic
The traffic dashboard is now actively misleading. Informational visits are migrating into AI answers, so sessions can fall while pipeline rises, and a programme judged on sessions will be cancelled in exactly the quarter it starts working. Measurement in 2026 runs on four levels. Pipeline attribution covers organic-sourced enquiries, their progression to opportunity and close, and revenue, pulled from the CRM rather than analytics. Conversion-layer performance means enquiries and demo bookings per commercial page, with AI-referral sessions (chatgpt.com, perplexity.ai, gemini) segmented in GA4 and watched through to enquiry rather than counted as sessions. Authority-layer performance tracks rankings on tier-1 and tier-2 terms, AI citation share across a fixed prompt set, and links and mentions earned. Hygiene is Search Console impressions, indexation and Core Web Vitals.
Two practices close the gaps the tools leave. First, self-reported attribution. Put "How did you hear about us?" on every form and log the answers against deals. Buyers increasingly do their research inside AI answers, then type your URL straight into the browser. Analytics calls that "direct". The buyer will happily tell you ChatGPT recommended them, and across a quarter those answers quantify the invisible channel better than any tracking pixel. Second, a monthly prompt-set run. Take the 20-40 questions your buyers ask and sample them across ChatGPT, AI Mode, Perplexity and Gemini. That gives you an AI share-of-voice figure to sit alongside your rank tracking.
Report the story in revenue language. Enquiries, opportunities, pipeline value and AI share of voice, trended, with traffic relegated to a diagnostic appendix. That framing is also the early-warning system. Falling informational traffic with stable enquiries means the funnel moved into the answer layer and your citations are carrying it. Falling enquiries means a real problem. Only pipeline-first measurement can tell those two apart.
The 6-month roadmap
Six months takes a typical B2B site from scattered blog posts to a working two-layer system. The sequence builds conversion capacity first (so authority traffic has somewhere to land), authority second, and treats measurement as a month-one deliverable rather than a year-end retrofit. Intensity flexes with budget. The order should not.
| Month | Focus | Deliverables |
|---|---|---|
| 1 | Foundations & baseline | Technical audit fixed (crawl, speed, schema, AI crawler access, Bing verification). Intent-tier keyword map from tools + CRM mining. Prompt-set baseline across 4 engines. Measurement stack live (GA4 segments, CRM source fields, "how did you hear about us") |
| 2 | Conversion layer I | Top service pages rebuilt answer-first with pricing ranges, FAQs, named authors. First competitor comparison page. Case studies structured with numbers |
| 3 | Conversion layer II + hub | Remaining service x market pages. Hub page for the head topic. Internal linking pass with descriptive anchors |
| 4 | Authority layer I | Pricing transparency guide and statistics page published ungated. Spokes for top tier-2 queries (cost, vs, how-to-choose) |
| 5 | Authority layer II + PR | Original research or survey piece. Digital PR push, trade press quotes, directory and list placements, regular LinkedIn expert publishing |
| 6 | Measure, prune, plan | Second prompt-set run vs baseline. Pipeline report (enquiries, opportunities, AI SOV). Content pruning and consolidation. Quarter-three plan weighted to what got cited and what converted |
Budgets, expectations and the honest timeline
UK SMEs should budget roughly £1,500 to £5,000 / $2,000 to $6,500 per month for a serious B2B SEO programme, with competitive sectors and multi-market coverage running £5,000 to £10,000 / $6,500 to $13,000. Within that, expect the spend to split approximately 30% technical and conversion-layer work (front-loaded), 50% authority content and digital PR (continuous), 20% measurement and iteration. Adding the GEO layer to an existing programme costs 15-30% more rather than double, because the foundations are shared. That is the strongest argument for buying both as one engagement.
Now the timeline. Months one to three are build months. Rankings move on long-tail and tier-1 terms, and the first AI citations appear on uncontested queries, but pipeline effects are early. Months four to six are when comparison pages start closing into deals, the authority layer earns its first steady citations, and the prompt-set trend becomes legible. A cited statistics page, a ranking pricing guide and a set of comparison pages then keep producing for years at marginal refresh cost. That is the opposite of how paid acquisition works, and it is why patient B2B firms end up being the ones AI answers recommend.
What B2B SEO costs in 2026
UK SME programmes run £1,500 to £5,000 / $2,000 to $6,500 per month. Competitive or multi-market programmes £5,000 to £10,000 / $6,500 to $13,000. One-off technical and content audits run £750 to £2,500 / $1,000 to $3,000. Expect conversion-layer effects from months 2-3, authority citations from months 3-4, and pipeline-level evidence by month 6. Blending GEO into the programme adds roughly 15-30%, sharing the same foundations.
Common questions
Is B2B SEO still worth it now AI answers half the queries?
Yes. Roughly a third of searches still send a click, bottom-funnel B2B queries retain strong click-through, and the visitors who arrive from AI answers convert better than ordinary organic ones. What changed is the strategy. Authority content wins the citations inside answers, conversion pages win the surviving high-intent clicks. Pure top-funnel blogging, by contrast, genuinely is dying.
How is this different from a GEO programme?
Same foundations, different emphasis. B2B SEO optimises for rankings, clicks and pipeline. GEO optimises for citations inside AI answers. The two-layer strategy in this guide serves both, and the efficient buy is one programme. The GEO layer adds roughly 15-30% to a competent SEO budget.
How long until B2B SEO shows pipeline results?
Tier-1 and long-tail rankings move within 2-3 months. Comparison and pricing pages typically produce their first attributable enquiries in months 3-4. Steady authority citations and pipeline evidence arrive around month 6. Anyone promising page-one rankings in weeks is describing either zero-competition terms or fiction.
Should we gate our whitepapers and research?
Mostly no. Gated PDFs cannot be crawled, cited or recommended. They earn nothing in the answer economy. Publish the citable substance (statistics, methodology, findings) as ungated HTML, and keep gating for the things a buyer will happily swap an email address for, such as a calculator or a template.
Do backlinks still matter for B2B?
For classic rankings, yes, authority remains part of the entry ticket. But unlinked brand mentions correlate with AI citations more strongly than backlinks do, and Reddit and LinkedIn are the most-cited domains in AI answers. Budget digital PR for mentions, reviews and expert visibility, and treat links as a welcome by-product.
Keep reading
Put this to work on your business
We build exactly this, SEO, AI visibility, PPC, email and automation, as one measured system. Talk it through with a real person, no obligation.
