SEO, GEO and AEO: How B2B Companies Get Found Online

  1. Blog
  2. Discoverability & Content

By Saroj Jha · January 7, 2026 · 11 min read

Buyers now find companies in three places: the ranked links on a results page, the answer boxes above those links, and answers written by AI systems such as ChatGPT, Perplexity and Google’s AI Mode. A page can do well in one and be missing from the other two.

The third place has grown fast. Google says AI Mode passed 1 billion monthly users in mid-2026, and it has merged AI Overviews and AI Mode into one experience[1]. ChatGPT has more than 900 million weekly users[2], and by mid-2025 about a quarter of its messages were people looking for information[3]. Three labels have grown up around this: SEO, AEO and GEO. They overlap more than the people selling them suggest, and Google treats all of it as SEO[4].

This post is the overview for a B2B company: what each label means, how the three fit together and what to do first. I link to the deeper AAJ articles on each part as I go.

SEO, AEO and GEO are three layers of one job. SEO gets your pages crawled, indexed and ranked; AEO makes them easy to lift as a direct answer; GEO makes your pages, and what other sites say about you, the sources AI engines retrieve and cite. Fix the shared foundation first, then answer the questions buyers ask, then make your facts match everywhere, and measure each layer on its own.

In three numbers

  • 1 billion monthly users of Google’s AI Mode by July 2026, Alphabet says[1]
  • 8% vs 15% of Google visits with a click on a regular result, with and without an AI summary (Pew, 2025)[5]
  • 2.2× the organic click-through of brands an AI Overview cites, against brands it leaves out: 2.07% to 0.94% (Seer, 2025)[6]

Three ways buyers find you, and what each one rewards

Each layer has its own surface and its own idea of winning, but all three stand on one foundation: pages a search engine can crawl, index and show with a snippet, that answer real questions with evidence.

Figure 1 SEO, AEO and GEO stand on one foundation and differ in what counts as winning
  • SEO Search engine optimization
  • AEO Answer engine optimization
  • GEO Generative engine optimization
Shared foundation Pages a search engine can crawl, index and show with a snippet, answering the questions buyers really ask, with evidence

Redrawn from the original article’s three cards, with GEO corrected from geographic to generative engine optimization (Aggarwal and colleagues, KDD 2024)[7]. Google treats all three as SEO[4]; its AI features use only pages that are indexed and eligible for a snippet[8].

A note on the name. GEO is sometimes used for geographic targeting. Here it means generative engine optimization, the term Aggarwal and colleagues introduced in a paper published at KDD 2024[7], and the meaning most of the industry now uses. Location still matters for a company that sells in particular markets, and I cover it under SEO.

The layers matter because clicks shrink where answers appear. In Pew Research Center’s panel of 900 US adults, people who saw a Google AI summary clicked a regular result on 8% of visits, against 15% when there was none, and clicked a link inside the summary on 1%[5]. Ahrefs found the top-ranked page drew 58% fewer clicks per impression than expected when an AI Overview appeared[9]. Being cited wins some of it back: across 53 brands in 2025, Seer Interactive measured organic click-through of 2.07% on informational queries where an AI Overview cited the brand, 0.94% where it didn’t and 3.35% where no AI Overview appeared[6]. These are correlations, two of them from vendors, but they point the same way: a ranking no longer guarantees the visit, and being the cited source is worth working for.

Figure 2 Fewer searchers click when an AI summary appears, and brands the summary cites keep more of the clicks

Share of Google search visits with a click Pew Research Center: 900 US adults, 68,879 searches, March 2025

  • No AI summaryClicked a regular result 15%
  • AI summary shownClicked a regular result 8%
  • AI summary shownClicked a link inside the summary 1%

Organic click-through on informational queries, 2025 Seer Interactive: 53 brands; vendor data

  • No AI Overview 3.35%
  • AI Overview cites the brand 2.07%
  • AI Overview, brand not cited 0.94%

Correlations, not causes. Gold bars are searches without an AI summary; blue bars are searches with one. Ahrefs, on 300,000 keywords, found the top-ranked page drew 58% fewer clicks per impression than expected when an AI Overview appeared.

Show the numbers as a table
Measure Value
Pew: no ai summary, clicked a regular result 15%
Pew: ai summary shown, clicked a regular result 8%
Pew: ai summary shown, clicked a link inside the summary 1%
Seer: No AI Overview 3.35%
Seer: AI Overview cites the brand 2.07%
Seer: AI Overview, brand not cited 0.94%
Ahrefs: top-ranked page, clicks per impression against forecast, AI Overview present −58%

Chapekis and Lieb, Pew Research Center, July 2025[5]; McDonald, Cooley and Williams, Seer Interactive, April 2026, full-year 2025 averages[6]; Law, Ahrefs, February 2026, December 2023 against December 2025[9].

Running case: Example Co. Fictional

Example Co. sells contract-management software to US construction subcontractors, such as electrical and drywall firms: it tracks contracts, change orders and pay applications. It is a 30-person Series A company with about 120 pages on its site. Search Console showed 18,000 impressions and 410 clicks a month, mostly for its own name. When the founder put 20 questions a buyer might ask, such as “best contract management software for subcontractors”, to ChatGPT and Perplexity, Example Co. was named in 2 of the 20 answers. All its details are illustrative.

Do this

Open Search Console and list the 20 queries that bring you the most impressions. Search each one in Google and in ChatGPT or Perplexity, and note which surface answers it and whether you appear. That one table shows which layer to fix first. For the strategy behind answers that never send a click, see Stop chasing the click.

SEO: the foundation all three share

SEO is the work of getting your pages crawled, indexed and ranked when people search for the problems you solve. Everything else in this post depends on it: Google’s AI features draw only on pages that are indexed and eligible to show with a snippet, and there are no extra technical requirements beyond that[8].

Start from real customer questions. Keyword tools show volume; sales calls, support tickets and meeting notes show the words buyers actually use. Recurring questions become pages that answer the question directly, show how you think and offer a clear next step. Example Co.’s sales team heard these every week:

  • “What does contract-management software cost for a 40-person electrical subcontractor?”
  • “Can we keep using our accounting system?”
  • “How do we prove a change order was approved?”

Build topic clusters, not isolated posts. A pillar page on the main problem, with supporting pages linked to it and to each other, helps both visitors and search engines see the depth of what you know.

Show who wrote it and why it can be trusted. Google says E-E-A-T (experience, expertise, authoritativeness and trust) isn’t a ranking factor in itself, that trust matters most, and it strongly encourages accurate bylines[10]. Name the author, show real examples and explain how you reached your numbers.

Show up in the markets you serve

Location still shapes results, and a company that sells in particular regions or countries should say so. Write a market page only when you have something specific to say there: customers, rules, partners, delivery or support in that market. Google’s spam policies list “pages targeted at specific regions or cities that funnel users to one page” as doorway abuse[11], so a page per city with only the name changed does more harm than good. If you sell in other languages or countries, use hreflang annotations, and don’t adapt content to the visitor’s IP address, which Google calls “generally not reliable”[12].

Check any advice to optimize a Google Business Profile. A profile requires in-person contact with customers during stated hours, and online-only businesses aren’t eligible[13]. If customers visit your office or showroom, keep the profile accurate; if not, put the effort into consistent listings on industry directories and partner pages. For what to do first on a young site, see SEO for Startups: When to Start and What to Do First.

Do this

Check the foundation before anything else. Verify your site in Search Console, confirm that your 20 most important pages are indexed, and read your robots.txt file and any bot-blocking setting at your hosting or CDN provider to see which crawlers they turn away.

AEO: be the answer a search engine can lift

AEO, answer engine optimization, means structuring pages so a search engine or assistant can lift a direct answer from them: a featured snippet, a “People also ask” box, a voice answer or a passage in an AI summary. The core move is simple. Put the exact question a buyer asks in a heading, and answer it in the next two or three sentences, in words that still make sense when quoted alone.

Example Co.’s answer, written to stand alone. “How do you prove a change order was approved? Keep the signed change order, or the written directive that authorized the work, with its date, scope and amount, attached to the contract it changes, in a form you can export the same day. Logging approvals as they happen closes the gaps phone calls leave.”

Cover the whole journey, not only definitions. Buyers ask different questions as they move from noticing a problem to choosing a supplier, and each stage needs pages of its own.

Figure 3 Each stage of the buyer’s journey asks different questions, and each needs a page of its own

Awareness Noticing the problem

  1. “Why do change orders go unpaid?” A guide that explains the causes
  2. “What does an unbilled change order cost a subcontractor?” A worked example with the arithmetic

Consideration Comparing ways to solve it

  1. “Contract software or spreadsheets?” A comparison page, with when each one fits
  2. “What should a subcontractor track on every contract?” A checklist of dates, amounts and approvals

Decision Choosing a supplier

  1. “What does it cost for a 40-person subcontractor?” A price-range page by company size
  2. “Can we keep using our accounting system?” An integrations and rollout page

Stages redrawn from the original article’s journey cards (awareness, consideration, decision). The questions are Example Co.’s and fictional.

Structured data has changed more than most AEO advice admits. Google stopped showing HowTo rich results in September 2023[14], and FAQ rich results stopped appearing on May 7, 2026[15]. FAQ markup does no harm, but it no longer earns a special result in Google. Google also says structured data isn’t required for its generative AI features, there is no special schema to add, and you don’t need to chop pages into small chunks[4]. What still pays is markup Google supports, such as Organization, Article, Product, Software app and Breadcrumb[16], kept consistent with the visible text.

Running case: Example Co. Fictional

Example Co.’s best-ranking page was a 2,000-word guide that answered its own title question in the ninth paragraph. The team moved the answer to the top, rewrote the headings of ten pages as the questions sales hears most, and added two pages buyers had asked for and never found: price ranges by company size, and how a rollout works.

Do this

Pick your ten most visited pages. For each, put the one question it answers in a heading near the top and answer it in two or three sentences that stand alone. Add the decision-stage questions sales hears most, such as cost, rollout and switching, as pages of their own. Leave existing FAQ markup in place, but don’t add FAQ or HowTo markup to win rich results, because Google no longer shows them.

GEO: be a source AI engines cite

GEO is the work of getting your pages, and what others say about you, retrieved and cited when an AI system writes an answer. It differs from SEO in two ways. The engine may run many searches for one question: Google says AI Overviews and AI Mode may issue “multiple related searches across subtopics and data sources” to build a response, which can surface a wider set of pages than classic results[8]. And it reads far more than your site, so reviews, directory listings, partner pages and press coverage count as much as your own copy.

The first GEO check is access, and it is the cheapest to fix. Each AI company runs separate crawlers for search and for training, and robots.txt treats them separately. Blocking OpenAI’s GPTBot keeps your content out of model training; blocking OAI-SearchBot keeps you out of ChatGPT search answers[17]. Anthropic and Perplexity split their crawlers the same way[18],[19]. Google-Extended controls whether Gemini models can train on or ground answers in your content, but not your inclusion or ranking in Google Search[20]. To keep a site out of AI Overviews and AI Mode, Google added a separate Search generative AI control in Search Console, open to all sites since August 31, 2026 and set to include by default[21].

Figure 4 Blocking a training crawler keeps you out of training; blocking a search crawler keeps you out of the answers
  • Company For search answers For model training When a user asks
  • OpenAI OAI-SearchBot: blocked sites aren’t shown in ChatGPT search answers GPTBot ChatGPT-User: robots.txt may not apply
  • Anthropic Claude-SearchBot: indexes content for search results ClaudeBot Claude-User: fetches pages for a user’s question
  • Perplexity PerplexityBot: not used for model training No separate crawler listed Perplexity-User: generally ignores robots.txt
  • Google Googlebot, plus the Search generative AI control in Search Console for AI Overviews and AI Mode Google-Extended: Gemini training and grounding; no effect on Search Not covered here

OpenAI, Overview of OpenAI Crawlers[17]; Anthropic Help Center[18]; Perplexity, Perplexity Crawlers[19]; Google, List of Google’s common crawlers[20] and Search Console Help[21]. OpenAI and Perplexity say robots.txt changes take about 24 hours to apply.

Two more technical points. In a 2024 analysis of Vercel’s network, none of the major AI crawlers it measured, including OpenAI’s, Anthropic’s and Perplexity’s, ran JavaScript; Google’s Gemini crawler did[22]. Content that appears only after scripts run may be invisible to them, so keep key text in the HTML. And llms.txt files are optional: Google says Google Search ignores them[4].

Once engines can reach you, the work is evidence and consistency. Answer questions directly, with figures and sources, and keep the basic facts the same on your site, your profiles, review sites and third-party listings: what you sell, who it is for, how it is priced and what it connects to. Research on what moves AI citations is younger and noisier than most GEO advice suggests; the widely quoted “40% visibility lift” came from a benchmark in which the pages were already in the model’s context[7]. I go through the evidence in What Is GEO?, and through the facts an engine needs before it will recommend you in Why doesn’t ChatGPT recommend my company?

Running case: Example Co. Fictional

A bot-blocking setting at Example Co.’s hosting provider turned away OAI-SearchBot, Claude-SearchBot and PerplexityBot along with the training crawlers. The founder allowed the search crawlers and kept GPTBot and ClaudeBot blocked, a deliberate choice about training, not about search. The team also found three different lists of supported accounting systems on its own site, its G2 profile and two partner listings, and made them one.

Do this

Write down which AI crawlers your robots.txt and CDN allow, and decide on purpose. Allow the search crawlers (OAI-SearchBot, Claude-SearchBot and PerplexityBot) unless you have a reason not to, and choose separately whether to allow the training crawlers. Then write a one-page fact sheet about the company and make every public profile match it.

Free tool

Check whether AI engines can crawl and cite a page

The free Crawl-to-Citation Scorer checks crawlability, server-rendered content, structured data and answer readiness for any URL, then lists the fixes in order.

Score a page

How the three fit together: one question, three surfaces

Take one question an Example Co. buyer might ask: “How do I stop losing money on change orders a general contractor approved only by phone?” The same work serves all three surfaces, but each rewards a different part of it.

Figure 5 One buyer question reaches three surfaces, and the same page can serve all three when each layer is in place

The buyer asks“How do I stop losing money on change orders a general contractor approved only by phone?”

  • Classic results SEO — The buyer sees A ranked guide, “How to get paid for change orders approved by phone” — Example Co. needed An indexed guide in the buyer’s words, linked from the change-order pillar page
  • Answer features AEO — The buyer sees A featured snippet or AI summary quoting the guide’s opening — Example Co. needed The question as a heading and a stand-alone answer: confirm in writing, record scope and amount, submit by the contract deadline
  • AI-written answer GEO — The buyer sees“Options include … Example Co., which logs field approvals from a phone”, with a citation — Example Co. needed Search crawlers allowed, the same facts on G2 and partner listings, a mention in a trade publication

Redrawn from the original article’s “How SEO, GEO and AEO work together” cards, with a new example. Example Co. and all its details are fictional and illustrative.

The shared work is the page itself. SEO gets it found and ranked, AEO makes its opening quotable, and GEO adds what the page can’t do alone: being reachable by AI crawlers and being confirmed by other sites. Treat them as three layers of one plan with one owner, not three projects.

Measure each layer separately

Being found only matters if it leads to pipeline, and each layer needs its own measure, because the three don’t move together.

Figure 6 Measure being found at three levels, from visibility to revenue
    • 1 Visibility
    • Search Console impressions, clicks and positions for your priority queries; generative AI report impressions; how often you are named across a fixed set of buyer prompts.
    • Is the right audience seeing you?
    • 2 Engagement and trust
    • Return visits to key pages; newsletter, assessment and tool sign-ups from organic search and AI referrals.
    • Do they come back?
    • 3 Pipeline and revenue
    • Qualified leads and opportunities from search and AI; “How did you hear about us?” answers; revenue from those deals.
    • Does it pay?

Redrawn and updated from the original article’s three levels of metrics. Google counts AI Overviews and AI Mode traffic in the Performance report under Web[8]; the generative AI performance reports count impressions, not clicks[23].

For Google, start in Search Console. Traffic from AI Overviews and AI Mode is counted in the Performance report under the Web search type[8], and the generative AI performance reports Google began rolling out in June 2026 show how often your pages appear in those features; they count impressions, not clicks[23]. For ChatGPT, Perplexity and Claude, run a fixed set of buyer prompts on a schedule and repeat each one several times, because the answers change from day to day; the free AI Visibility Significance Test tells you whether a change is real.

Then ask every new lead how they found you. Answers such as “ChatGPT suggested you” or “you kept coming up when I searched” catch what no dashboard sees, so add “How did you hear about us?” as an open text field on your demo form. For the tools that support this tracking, see the SEO and content layer of the 2025–2026 MarTech landscape report.

Running case: Example Co. Fictional

Ninety days later, Example Co. was named in an average of 7 of the same 20 prompts across five runs, against 2 before, and Search Console’s generative AI report showed 14 of its pages appearing as sources in AI Overviews and AI Mode. Monthly organic clicks rose from 410 to 520. Six of the quarter’s 45 demo requests named ChatGPT or Perplexity in the “How did you hear about us?” field. Example Co. can’t say which change did the most; it can say that all three layers moved.

Do this

Set a baseline before you change anything. Record Search Console clicks and impressions for your 20 priority queries, run a frozen set of 20 buyer prompts five times in two AI engines, and add an open “How did you hear about us?” field to your demo form. Review all three once a month.

What to do first: a 30-day order

If you do nothing else, do these in order. Each step makes the next one count.

Figure 7 A 30-day order: foundation, questions, answers, then facts and a baseline
  1. Week 1 Foundation. Verify Search Console and check that your 20 key pages are indexed. Open robots.txt and CDN settings to the search crawlers, and keep key text in the HTML.
  2. Week 2 Questions. Collect 20 buyer questions from calls, tickets and emails. Sort them into awareness, consideration and decision, and pick the ten pages to fix.
  3. Week 3 Answers. Put each question in a heading with a stand-alone answer. Add named authors and sources, and write the missing decision-stage pages.
  4. Week 4 Facts and baseline. Write one fact sheet and match every profile and listing to it. Record 20 queries in Search Console, run 20 prompts five times, add a “How did you hear about us?” field.

My recommended order, built from the Do this steps in this article.

None of this needs a large team. It needs one person who owns the questions buyers ask and makes sure the answers are findable, quotable and the same everywhere.

Related on AAJ

Articles

Playbooks

Reports

Templates and tools

References

The superscript numbers in the article match this list.

Studies and web pages were checked on October 21, 2026.

  1. Pichai, S. (2026). “Q2 2026 earnings call: Remarks from our CEO.” Google, The Keyword, July 22, 2026. AI Mode passed 1 billion monthly active users; AI Overviews and AI Mode brought together into one Search experience. blog.google
  2. OpenAI (2026). “Scaling AI for everyone.” February 27, 2026. More than 900 million weekly active ChatGPT users. openai.com
  3. Chatterji, A., Cunningham, T., Deming, D. J., Hitzig, Z., Ong, C., Shan, C. Y. and Wadman, K. (2025). “How People Use ChatGPT.” NBER Working Paper 34255, September 2025. Not peer reviewed. About 1.1 million sampled consumer conversations, May 2024 to June 2025; “Seeking Information” rose from 14% to 24% of messages over the year. nber.org
  4. Google Search Central (2026). “Optimizing your website for generative AI features on Google Search.” Published May 15, 2026; last updated July 10, 2026. developers.google.com
  5. Chapekis, A. and Lieb, A. (2025). “Google users are less likely to click on links when an AI summary appears in the results.” Pew Research Center, July 22, 2025. Browsing data from 900 US adults in KnowledgePanel Digital; 68,879 Google searches, March 1 to 31, 2025. pewresearch.org
  6. McDonald, T., Cooley, H. and Williams, M. (2026). “AIO Impact on Google CTR: 2026 Update.” Seer Interactive, April 24, 2026. Vendor study: 53 brands, 5.47 million tracked queries, January 2025 to February 2026; full-year 2025 averages for informational queries. seerinteractive.com
  7. Aggarwal, P., Murahari, V., Rajpurohit, T., Kalyan, A., Narasimhan, K. and Deshpande, A. (2024). “GEO: Generative Engine Optimization.” Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2024). Peer reviewed. dl.acm.org
  8. Google Search Central (2025). “AI features and your website.” Last updated December 10, 2025. developers.google.com
  9. Law, R. (2026). “Update: AI Overviews Reduce Clicks by 58%.” Ahrefs, February 4, 2026. Vendor study: 300,000 keywords; aggregated desktop Google Search Console click-through for the top-ranked page, December 2023 against December 2025, compared with a forecast from keywords without AI Overviews. ahrefs.com
  10. Google Search Central (2026). “Creating helpful, reliable, people-first content.” Last updated October 5, 2026. developers.google.com
  11. Google Search Central (2026). “Spam policies for Google web search.” Doorway abuse and scaled content abuse. Last updated August 28, 2026. developers.google.com
  12. Google Search Central (2025). “Managing multi-regional and multilingual sites.” Last updated December 10, 2025. developers.google.com
  13. Google (2026). “Business eligibility and ownership guidelines.” Google Business Profile Help. support.google.com
  14. Google Search Central Blog (2023). “Changes to HowTo and FAQ rich results.” August 2023; HowTo rich results stopped showing on desktop on September 13, 2023, after mobile. developers.google.com
  15. Google Search Central (2026). “Latest documentation updates.” Entries of May 8, 2026 (FAQ rich results no longer shown from May 7, 2026) and June 15, 2026 (FAQ rich result documentation removed). developers.google.com
  16. Google Search Central (2026). “Structured data markup that Google Search supports.” Last updated October 16, 2026. developers.google.com
  17. OpenAI (2026). “Overview of OpenAI Crawlers.” OAI-SearchBot, GPTBot, ChatGPT-User and OAI-AdsBot. developers.openai.com
  18. Anthropic (2026). “Does Anthropic crawl data from the web, and how can site owners block the crawler?” ClaudeBot, Claude-User and Claude-SearchBot. support.claude.com
  19. Perplexity (2026). “Perplexity Crawlers.” PerplexityBot and Perplexity-User. docs.perplexity.ai
  20. Google (2026). “List of Google’s common crawlers.” Google-Extended. Last updated October 15, 2026. developers.google.com
  21. Google (2026). “Search generative AI control.” Search Console Help. Available for all websites from August 31, 2026. support.google.com
  22. Zecchini, G., Moore, A. A., Ubl, M. and Siddle, R. (2024). “The rise of the AI crawler.” Vercel, December 17, 2024, with MERJ. Crawler traffic on nextjs.org and the Vercel network. vercel.com
  23. Maoz, H. and Samet, M. (2026). “Introducing Search Generative AI performance reports in Search Console.” Google Search Central Blog, June 3, 2026. developers.google.com

More in Content, Copy & SEO

Part of the Content, Copy & SEO hub - see all 20 resources on this topic.

Also useful in AI Search & Agent Readiness.