7 Generative Engine Optimization Fixes: Complete 2026 Guide
Seven generative engine optimization fixes for a B2B website, with the KDD study results, a problem filter and a next-fix checker.

Published 20 April 2023 · Last updated and fact-checked 28 October 2026
Short answer: Generative engine optimization comes down to seven fixes on a site that already exists: let the AI crawlers in, write one facts block, put numbers and causes in claims, cite sources, answer first, mark up the facts, and measure. In the original study, adding citations raised a fifth-ranked site’s share of the answer by 115.1% (Aggarwal et al., KDD, 2024).
Generative engine optimization (GEO) is the work of getting your company named in the answer when a buyer asks ChatGPT, Claude, Perplexity or Google’s AI Overviews a question, and quoted closely enough that a link back to your site follows.
A head of operations at a 60-person engineering firm types “who should rebuild our HubSpot website” into ChatGPT. The answer names three firms. Your firm ranks on page one of Google for that exact search. It is not in the answer. Same site, same rankings, different answer.
I run Jackai Agency from Vancouver, where I build B2B sites mostly on WordPress. This guide covers what generative engine optimization is, the seven fixes I make on a B2B website so AI search tools can use it, and how I measure whether any of it worked. I also sell this work, so weigh my advice the way you would weigh any vendor’s.
In this guide
Key takeaways
- GEO sits on top of SEO. In the original study the engine read only the top five Google results, so a page the engine never fetches cannot be quoted.
- Citing sources on the page raised a fifth-ranked site’s share of the answer by 115.1% and cut the top-ranked site’s share by 30.3%.
- Keyword stuffing gave little to no lift, and a 10% loss on Perplexity.
- Only 51.5% of sentences from four generative search engines were fully supported by their citations, so read what each answer says about you.
- Forrester reported in January 2026 that 94% of business buyers use AI somewhere in their buying process.
| Fix | What changes on your site | Best for | Effort (my estimate) |
|---|---|---|---|
| 1. Let the AI crawlers in | robots.txt, security plugin, CDN rules | Sites blocking crawlers by accident | 1 hour |
| 2. Write one facts block | Same company facts in every place | Firms described three ways online | 1 day |
| 3. Put numbers and causes in claims | Service page and case study sentences | Pages full of vague claims | 1 week |
| 4. Cite sources on the page | Linked studies inside the paragraph | Pages that make buyer-behavior claims | 2 days |
| 5. Answer first on every page | The first line of each service page | Pages that open with a mission statement | 1 week |
| 6. Mark up the facts | Schema.org markup, optional llms.txt | Sites with no structured data | 2 days |
| 7. Measure it three ways | Recorded answers, AI referral traffic, leads with source | Anyone about to spend money | 2 weeks |
What is generative engine optimization?
Generative engine optimization is the practice of making a page the source an AI search answer is built from, with your name attached. A search engine gives you a place on a list. A generative engine gives you a share of one answer, or nothing, and the link inside that answer is where AI referral traffic starts.
The term comes from a paper by Pranjal Aggarwal and five co-authors, first posted in 2023 and published in 2024 in the proceedings of KDD, a peer-reviewed data mining conference (Aggarwal et al., 2024). They describe a generative engine as a system that fetches sources for a question and then has a language model write one answer from them, with citations. Their examples were Bing Chat, Google’s Search Generative Experience (now AI Overviews) and Perplexity.
To measure visibility they counted how many words of the response came from your page, weighted by how early those words appeared. They called it position-adjusted word count. The authors, from the Indian Institute of Technology Delhi and Princeton University among others, summarize the result in one line:
“GEO can boost visibility by up to 40% in generative engine responses,” wrote Aggarwal and co-authors (Aggarwal et al., 2024, Indian Institute of Technology Delhi and Princeton University).
I find that a useful way to set the goal: you are not trying to rank, you are trying to be the source the answer is built from. Answer engine optimization (AEO) is the close cousin: formatting a single page so a clean answer can be lifted from it. I treat the two as one job with two layers. GEO decides whether an engine uses your site at all. AEO decides how cleanly it can quote a given page. On my AI Visibility Optimization service they are one six-week engagement.
How is generative engine optimization different from SEO?
GEO sits on top of SEO, not in place of it. Rankings still decide who gets into the pool of pages an engine reads, and what changes is what wins once you are in it. A search results page rewards relevance, links and speed; an AI answer rewards quotable facts with numbers and named sources.
In the paper’s test setup, the engine read only the top five Google results for each query before it wrote anything. Real AI search tools retrieve in their own ways, but the rule holds: a page the engine never fetches cannot be quoted.
Keyword stuffing, the oldest SEO habit, gave little to no lift in the engine the authors built. On Perplexity it did 10% worse than leaving the page alone. Adding citations to credible sources did the opposite, and the gains landed where SEO helps least. For the site ranked fifth in search results, citing sources raised its share of the answer by 115.1%. The same change on the top-ranked site cut its share by 30.3%.
| SEO | GEO | |
|---|---|---|
| What you win | A position on a results page | Words and a citation inside an AI search answer |
| Who reads first | A person scanning ten links | A language model reading a handful of sources |
| What moves it | Relevance, links, page speed, crawl health | Relevance first, then quotable facts with numbers and named sources |
| What fails | Thin or slow pages | Vague claims, facts that conflict across sites, keyword stuffing |
| How you measure | Rank, clicks, Search Console | Recorded answers, AI referral traffic, leads by source |
That fifth-place result is the part of generative engine optimization that interests me most for smaller B2B firms: once your page is in the pool, the model reads your words, not your backlink count.
1. Let the AI crawlers in: best for sites that block them by accident
Check robots.txt, the security plugin and the CDN firewall for rules that block the crawlers AI search tools publish. If the crawler is blocked, there is no answer to appear in and no AI referral traffic to count. It is the cheapest fix on this list, so it goes first.

- Fixes: a site invisible to AI search tools
- Where to look: robots.txt, security plugin, CDN firewall rules
- Crawlers named by their makers: OpenAI’s GPTBot and OAI-SearchBot, PerplexityBot, ClaudeBot
- Effort: about 1 hour (my estimate)
A blocking rule can be switched on without anyone in marketing knowing. A security plugin updates, a firewall preset treats unfamiliar bots as hostile, and a developer adds a disallow line during a staging launch and forgets it. Nobody notices because Google still crawls the site and rankings look normal.
The test takes minutes. Open your robots.txt, search for each crawler name, and then check the firewall log for blocked requests from those user agents. If a rule is there, remove it or scope it to the pages you want kept private, such as a staging folder or a client portal.
The dull answer is the useful one here. You do not need a new tool or a new page. You need to confirm that the engines your buyers use are allowed to read the pages you already wrote, and that is a precondition for every other fix in this guide.
Data point: in the KDD study the engine read only the top five Google results for each query, which shows how much depends on being fetched at all (Aggarwal et al., 2024).
2. Write one facts block: best for firms described three ways online
A facts block is one short set of statements about your company, written once and repeated word for word across the About page, footer, LinkedIn and directory profiles. Engines check you against other sites. When your site and your profiles describe you three ways, the engine picks one, and you do not get to choose which.

- Fixes: conflicting descriptions of what you sell
- Contents: company name, what you sell, who it is for, where you work, a starting price
- Where it lives: About page, footer, LinkedIn, every directory profile
- Effort: about 1 day (my estimate)
My own About page has a table called “The facts, in one place” for exactly this. Each row is a sentence a machine can lift whole: who runs the agency, where it is based, what it sells and what a project starts at.
The work is mostly editing, and it is tedious in the right way. Collect every place your firm describes itself, paste the descriptions into one document, and mark where they disagree. Common splits: a service called “web design” on the site and “digital experience” on LinkedIn, a team of 12 on one page and 20 on another, a founding year that differs by two.
Pick the version that is true today and change the rest to match. Where you control the wording, use identical words. Where you do not, such as a review site or an old press listing, ask for a correction and note which ones you could not fix.
Data point: language models weigh conflicting web pages mainly by how relevant each is to the question, so a clear and consistent page has to win on relevance first (Wan, Wallace and Klein, 2024).
3. Put numbers and causes in claims: best for pages full of vague claims
“We grow organic traffic” gives an engine nothing to lift. A claim an engine can quote names who, what moved and why. The KDD paper found that adding statistics in place of vague description was one of the three biggest gains, each raising position-adjusted word count by 30 to 40%.

- Fixes: claims too vague to quote
- Pattern: who, what moved, why, and the caveat in the same sentence
- Proof I can publish: Advisor Websites and Snappy Kraken
- Effort: about 1 week for the main service pages (my estimate)
Compare: Advisor Websites, a Vancouver SaaS platform for financial advisors, saw +58% organic traffic and +28% MQL conversion from the replay library and opt-in templates. That sentence names who, what moved and why. It can be quoted whole.
Keep the caveat in the sentence. On Snappy Kraken, a FinTech SaaS for financial advisors, the redesign influenced 7 figures of ARR, influenced and not attributed, because sales, pricing and product moved in the same period. If the caveat sits two paragraphs later, a machine that lifts one sentence lifts the claim without it.
Go through each service page and mark every sentence that claims a result. For each, ask three questions: can a stranger tell who it happened to, is there a number, and does the sentence say why. If you cannot support a number, do not invent one. Say what you do know, such as the method, and leave the result out.
“If the caveat sits two paragraphs later, a machine that lifts one sentence lifts the claim without it,” wrote Joshua Jackai, Vancouver web designer.
“Plain and specific beats confident and vague,” wrote Joshua Jackai, Vancouver web designer.
Data point: the three best rewrites in the KDD paper, citations, quotations and statistics, each raised position-adjusted word count by 30 to 40% over the unchanged page (Aggarwal et al., 2024).
4. Cite sources on the page: best for pages that make buyer-behavior claims
When a service page makes a claim about buyer behavior, the study goes in the same paragraph, linked. In the KDD tests, citing credible sources was among the three changes with the biggest gains, and it kept the writer honest as a side effect.

- Fixes: claims that read as opinion
- Rule: the study sits in the same paragraph as the claim, with a link
- Result in the paper: +115.1% for the fifth-ranked site, -30.3% for the top-ranked site
- Effort: about 2 days (my estimate)
The result deserves a careful reading. For the site ranked fifth in search results, citing sources raised its share of the answer by 115.1%. The same change on the top-ranked site cut its share by 30.3%. Same page, same rank, a bigger slice of the answer for the page that had further to climb.
A second study adds a check. When language models weigh conflicting web pages, they rely on relevance to the question.
“current models rely heavily on the relevance of a website to the query, while largely ignoring stylistic features that humans find important,” wrote Wan, Wallace and Klein (Wan, Wallace and Klein, 2024, Boston College and University of California, Berkeley).
That does not cancel the citation result. It sets the order: relevance to the buyer’s actual question comes first, and sources help a relevant page, not an off-topic one.
Do the work in order, and keep it small. Pick the five pages that answer your buyers’ most expensive questions, find one or two peer-reviewed or primary sources for each claim, and link them inline. Label vendor data as vendor data. Skip the claim if you cannot find a source.
5. Answer first on every service page: best for pages that open with a mission statement
The first line of a service page should answer the one buyer question the page exists for. Everything after it backs that answer up. Language models use information best at the start or the end of what they read, so the opening line is the part most likely to be used.

- Fixes: buried answers and mission-statement openers
- Rule: one buyer question, answered in the first 40 to 60 words
- Supporting study: Liu et al., 2024, Transactions of the ACL
- Effort: about 1 week for the main service pages (my estimate)
When researchers moved the key passage around inside a long input, language models used it best at the start or the end, and noticeably worse in the middle.
“performance is often highest when relevant information occurs at the beginning or end of the input context,” wrote Liu and co-authors (Liu et al., 2024, Stanford University and University of California, Berkeley).
I cannot prove every engine reads a web page that way, but it is one more reason to put the answer first.
A mission-statement opener is the usual failure: “We help ambitious companies grow.” It answers nothing. Replace it with the question a buyer typed to find you, answered flat. For a page about a WordPress rebuild, that is the price, the timeline and who it suits, in two sentences. Then give the detail.
Write the question as an H2 or H3 in the buyer’s own words where it fits, and follow it with a 40 to 60 word answer. That block can be lifted as it stands, which is what the engine is looking for.
Data point: the same Stanford and Berkeley study ran two tasks, multi-document question answering and key-value retrieval, and found the position effect in both (Liu et al., 2024).
6. Mark up the facts, then decide on llms.txt: best for sites with no structured data
Add Schema.org structured data for the organization, each service and each offer. It does not make an engine cite you, but it removes the guessing about what you sell and what it costs. An llms.txt file is optional and I treat it as low priority.

- Fixes: machines guessing what you sell and what it costs
- Markup: Organization, Service, Offer, FAQPage where the page has real FAQs
- Standard: created by Google, Microsoft and others, with 31.3% of sampled pages using it by 2015
- Effort: about 2 days (my estimate)
Schema.org is a shared vocabulary that Bing, Google and Yahoo created in 2011, with Yandex joining later, so one set of markup could be read by all of them. By 2015, 31.3% of pages in a 10-billion-page sample carried it (Guha, Brickley and Macbeth, 2016). It is old and dull, which is what you want from a standard.
Mark up only what is on the page. If a service has a starting price, put the same price in the Offer. If the page has no FAQ, do not add FAQPage markup to get one. Then test the output in a validator and fix errors before moving on.
I also add an llms.txt file, a plain-text summary of the business at the site root. It takes about an hour, and I would not pay anyone for it on its own: I know of no study showing it changes what an engine says. Engines also check you against other sites: directories, reviews, articles that mention you. The facts block from fix 2 should match every one of them.
“It is old and dull, which is what you want from a standard,” wrote Joshua Jackai, Vancouver web designer.
Data point: Guha, Brickley and Macbeth describe Schema.org as a shared vocabulary for structured data, with adoption measured in a sample of 10 billion pages (Guha et al., 2016, industry practice article, not peer reviewed).
7. Measure it three ways: best for anyone about to spend money
Generative engine optimization gets measured with three numbers, each from a different place: what the answers say, AI referral traffic split out of direct, and leads with a source attached. Any one of them alone will mislead you. If you can only wire up one, wire up the third.

- Fixes: not knowing whether any of this worked
- Numbers: recorded answers, AI referral traffic, leads by source
- Cadence: the same questions, run again and dated
- Effort: about 2 weeks to set up (my estimate)
What the answers say. Write down 10 to 20 questions your buyers actually type. Run each one several times in each AI search tool, record the answers word for word, and date them. On my offers page I show one real check from 3 October 2026. The question was “Who is the best web design agency in Vancouver?” Three agencies were each named in 3 of 3 ChatGPT answers. The names are hidden, because the count is the point.
Read what each answer says about you, not only whether you are linked. A 2023 audit of four generative search engines (Bing Chat, NeevaAI, Perplexity and YouChat) found that only 51.5% of generated sentences were fully supported by their citations, and only 74.5% of citations supported the sentence beside them (Liu, Zhang and Liang, 2023).
“frequently contain unsupported statements and inaccurate citations,” wrote Liu, Zhang and Liang (Liu, Zhang and Liang, 2023, Stanford University).
That audit did not cover ChatGPT, but I check ChatGPT citations for the same failure: a link to your page beside a sentence your page never said. Track that as a fix, not a win. Count linked citations apart from mentions without a link. A mention can put you on the shortlist, but only a link sends a visit you can measure.
AI referral traffic, split out of direct. Without a rule for it, AI referral traffic either lands in the direct bucket, next to people typing your URL, or hides among ordinary referral sites. The method on my lead attribution case study splits it out. AI tools that pass a referrer are matched and given their own channel. Visits that arrive with no referrer are inferred from the landing page and session pattern, and reported as probable rather than certain.
Expect AI referral traffic to undercount. The research firm Forrester reported in January 2026 that 61% of business buyers use private AI tools provided by their own company. Research done behind a firewall may never send a referrer at all. That is why the recorded answers matter even when the traffic line looks flat.
Leads with a source attached. Hidden fields on every form carry the original source, the referring site and the landing page into the CRM with the lead. The weekly report then gives AI its own row, so ChatGPT citations and Claude referrals sit next to organic, paid and email. On the case study the cells are empty on purpose: client figures from attribution engagements are not published, so the page shows the format, not a result.
“A mention can put you on the shortlist, but only a link sends a visit you can measure,” wrote Joshua Jackai, Vancouver web designer.
Recorded answers tell you whether you are named. AI referral traffic tells you whether anyone clicked. Leads with a source tell you whether it paid.
What does the research say, and where does it stop?
The research supports a direction, not a forecast. The KDD paper tested nine rewrites on a benchmark of 10,000 queries, and the limits matter as much as the gains. Treat the numbers as shares of an answer in a lab setup, not as traffic or leads for your site.
The three best rewrites were adding citations to credible sources, adding quotations, and adding statistics in place of vague description. Two quieter changes also worked: making the text more fluent, or simply easier to read, gave gains of 15 to 30%. One change did not work: rewriting a page in a more persuasive, authoritative tone gave no significant improvement. The engines were not impressed by confidence.
Now the limits. The KDD tests ran on an engine built on GPT-3.5 and on Perplexity, a live AI search engine, as both stood when the paper was written, and engines have changed since. I use the paper as a direction for what to write, not as a forecast for any one site.
What I would not do
Some generative engine optimization advice in circulation is SEO spam with a new label. These are the moves I turn down, even when a client asks. Each one either fails in the research or puts a client’s name on a trick that I expect to age badly.
- Hide instructions for the AI. A 2024 study showed that hidden text written to steer the language model could push low-ranked products up in conversational search results, and the attack carried over to Perplexity (Pfrommer et al., 2024). It is the same trick as hidden keyword text in the early 2000s. I expect it to end the same way.
The authors, at the University of California, Berkeley, put the transfer result plainly:
“these attacks transfer effectively to state-of-the-art conversational search engines such as perplexity.ai,” wrote Pfrommer and co-authors (Pfrommer et al., 2024, University of California, Berkeley).
The other moves I turn down are shorter to explain:
- Stuff the keyword. The KDD paper found little to no gain in its own engine, and a 10% loss on Perplexity.
- Publish 200 AI-written pages. Part of the KDD impression score is how unique a source’s material is. A page that repeats what ten others say gives the engine no reason to pick it.
- Buy a guarantee. Nobody controls what ChatGPT says, so nobody can sell you ChatGPT citations. My own offers page says it plainly: no citation or ranking is guaranteed.
- Rewrite everything louder. The authoritative-tone rewrite gave no significant gain. Plain and specific beats confident and vague.
Compare the seven fixes and filter by problem
The seven fixes differ in cost and in what they repair. This matrix lets you filter by the problem you have and the time you can spend, so you can choose a first move. Effort figures are my estimates from client work, not benchmarks.
| Fix | Problem it repairs | Effort (my estimate) | You hold |
|---|---|---|---|
| 1. Let the AI crawlers in | Crawlers blocked without anyone knowing | 1 hour | A clean robots.txt and firewall log |
| 2. Write one facts block | Three descriptions of the same firm | 1 day | One block, repeated word for word |
| 3. Put numbers and causes in claims | Claims no one can quote | 1 week | Who, what moved, why, caveat |
| 4. Cite sources on the page | Buyer-behavior claims with no study | 2 days | Linked studies inside paragraphs |
| 5. Answer first on every page | Mission-statement openers | 1 week | A 40 to 60 word answer on top |
| 6. Mark up the facts | Machines guessing what you sell | 2 days | Valid Schema.org markup |
| 7. Measure it three ways | No read on what worked | 2 weeks | Dated answers, AI traffic row, lead sources |
Check which fixes your site already has
A quick self-check beats a long audit. Tick the fixes already in place on your site, and the tool names the first one still missing, in the order I would do them. It uses only your answers, and it is a starting point, not a score.
Which fix is next for your site? Tick what is already in place. Nothing is stored or sent.
How long does generative engine optimization take, and what does it cost?
My AI Visibility Optimization engagement takes six weeks and costs CA$4,775, on the website you already have. That covers fixes 1 to 7 on the pages that matter most, plus a dated baseline and a week-six comparison. Pricing for other work is on the offers page.
- A dated baseline of what each engine says about you, across your buyer questions.
- Schema.org markup for the organization, every service and every offer.
- Service pages rewritten so each answers one buyer question in its first line.
- FAQ content matching real buyer queries, marked up as FAQPage.
- An llms.txt file at the site root.
- A citation package of the directory and profile listings the engines cite in your category.
- A week-six comparison: the same questions, run again against the same baseline.
On timing, be realistic. Engines that search the web live can use a changed page once they crawl it again. What a model learned in training changes only when the model is retrained. So a six-week window mostly shows the live-search side. AI search answers also shift from week to week with nothing changed on your side, which is why every baseline is dated.
Every quote is fixed before work starts, with two revision rounds per phase, and you own the site, domain and hosting. If you cannot yet tell where leads come from, Lead Source Reporting separates AI referral traffic from direct, takes two weeks and costs CA$2,600. Without it, you will not know whether generative engine optimization paid off.
When should a B2B firm hold off?
Hold off on paid work until you know what the engines say today. Forrester reported in January 2026 that 94% of business buyers now use AI somewhere in their buying process, so the question is rarely whether buyers use these tools. It is what to fix first, and this table maps your situation to a starting point.
| Your situation | Start with | Why |
|---|---|---|
| You do not know if AI search tools name you | Free AI Visibility Report, 24 hours | A dated baseline before anyone spends money |
| You rank in Google, but competitors get named in AI search answers | AI Visibility Optimization, CA$4,775, six weeks | The site is found but not quoted |
| You cannot say where last quarter’s leads came from | Lead Source Reporting, CA$2,600, two weeks | Without it, you cannot tell whether GEO worked |
| Your site cannot say what you do in one sentence | Website Audit, CA$995, credited in full against a build (current terms on the offers page) | The fix is the site, not the markup |
| You are named and want to stay named | AI Referral Retainer, CA$1,870 a month, 90-day minimum | Answers change with nothing changed on your side, so someone re-runs the questions |
One edge case matters in professional services: your best proof may be under NDA. An engine can only quote what is public. In that case, publish the method instead of the result. My lead attribution page does exactly that.
The other edge case is a site with no clear service pages at all.
“No amount of generative engine optimization helps a page that does not exist,” wrote Joshua Jackai, Vancouver web designer.
Fix the structure first, then mark it up.
Generative engine optimization questions answered
What is generative engine optimization in simple terms?
Generative engine optimization is editing your site so AI search tools can quote it and link back. The term comes from a 2024 KDD paper, which found that citing sources, adding quotations and adding statistics each raised a page’s share of an AI answer by 30 to 40%.
Is generative engine optimization different from SEO?
Yes, but it builds on SEO. Rankings decide whether an engine reads your page at all, since the study engine read only the top five Google results. After that, quotable facts with numbers and named sources win, and keyword stuffing gives little to no lift.
Can anyone guarantee ChatGPT will cite my company?
No. Nobody controls what ChatGPT says, so a guarantee is a sales line. A 2023 audit found only 51.5% of sentences from four generative search engines were fully supported by their citations. What you can do is make your pages easy to quote and measure the result.
How do I measure generative engine optimization?
Use three numbers: recorded answers to 10 to 20 buyer questions, dated and repeated; AI referral traffic split out of direct; and leads with their original source attached. If you can only wire up one, pick leads with a source, because it shows whether the work paid.
Find out what AI search says about you today
Not a redesign and not a ranking report. Before you spend anything on generative engine optimization, get a dated record of what ChatGPT, Claude and Google AI Overviews say when your buyers ask. Same site, same rankings, one new baseline, so every later change has something honest to be compared with.
Send me the website and a work email, and I will send back the free AI Visibility Report in 24 hours: the questions, the answers, who gets named in your place, what the engines could not find on your site, and three fixes in order. No sales call unless you book one.
Sources
Peer-reviewed research
- 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. https://doi.org/10.1145/3637528.3671900
- Liu, N., Zhang, T. and Liang, P. (2023). Evaluating Verifiability in Generative Search Engines. Findings of the Association for Computational Linguistics: EMNLP 2023. https://doi.org/10.18653/v1/2023.findings-emnlp.467
- Wan, A., Wallace, E. and Klein, D. (2024). What Evidence Do Language Models Find Convincing? Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics. https://doi.org/10.18653/v1/2024.acl-long.403
- Liu, N. F., Lin, K., Hewitt, J., Paranjape, A., Bevilacqua, M., Petroni, F. and Liang, P. (2024). Lost in the Middle: How Language Models Use Long Contexts. Transactions of the Association for Computational Linguistics. https://doi.org/10.1162/tacl_a_00638
- Pfrommer, S., Bai, Y., Gautam, T. and Sojoudi, S. (2024). Ranking Manipulation for Conversational Search Engines. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. https://doi.org/10.18653/v1/2024.emnlp-main.534
Industry and vendor data (not peer reviewed)
- Guha, R. V., Brickley, D. and Macbeth, S. (2016). Schema.org: Evolution of Structured Data on the Web. Communications of the ACM, practice article by engineers from the companies behind Schema.org. https://doi.org/10.1145/2844544
- Buten, J. (22 January 2026). B2B Buyers Make Zero-Click Number One. Forrester blog, reporting Forrester’s Buyers’ Journey Survey.
- Jackai Agency client figures (Advisor Websites, Snappy Kraken): first-party results, not independently audited.


