Skip to content
Chalk Labs / AI SEO
Industry guide

AI SEO for Crypto and Web3: How to Get Recommended by ChatGPT and Google AI Overviews

By Rahil Jain, founder of Chalk LabsUpdated How we research

The short answer

A Web3 project gets recommended when its documentation, research and independent coverage let AI check what the project does, what it cannot do and who stands behind it. In crypto, claims about security, privacy and returns carry risk, so precise, dated, sourced wording beats hype. Publish docs that state limits, and earn real coverage rather than bought links.

  • 84% of developers in Stack Overflow's 2025 survey use or plan to use AI tools, and 66% name 'almost right, but not quite' as the most common frustration. 1
  • A voting-privacy question mixes layers: tool, cryptography and chain. Public docs for each state different limits, so say yours plainly. 23
  • Official sites supplied 63.4% of ChatGPT citations on factual questions but 46.4% on comparisons, so docs plus outside coverage both matter. 6
  • Paid media placements can break Google's site reputation policy or SEC and UK promotion rules if handled badly. 121314
  • Chalk Labs has no published AI-citation results for Web3 clients, and we found no public study that isolates crypto. 16
Why it matters in Crypto and Web3
  • Builders use AI heavily: 84% of developers use or plan to use AI tools 1.
  • 'Almost right' answers are costly for contracts, addresses and security, and 66% of developers name them as their top frustration 1.
  • Comparison prompts lean on outside sources: official sites got 46.4% of ChatGPT citations on comparisons against 63.4% on facts 6.
  • Paid promotion and bought placements carry rules and policy risk 121314.

What buyers in Crypto and Web3 ask AI

Pick a question to see what the buyer wants and what to publish so an AI can cite you.

  • What they want
    Choose a technology for secret ballots
    What to publish
    A page that separates the voting tool, the privacy method and the chain, and states each one's limits

    Searches we expect (modelled by Chalk Labs, not observed)

    • private voting DAO
    • shielded voting DAO
    • MACI vs shielded voting
    • DAO voting privacy trade-offs

What could developer and partner enquiries be worth?

Enter your own numbers to see how visits to your docs and decision pages turn into builder or partner enquiries, signed integrations and revenue each month.

People who arrive from Google, ChatGPT, Perplexity and similar.
Form fills, calls and emails, as a share of visits.

What this adds up to

About 6 customers a month and $48,000 in revenue. After a monthly cost of $2,500, that leaves $45,500.

About 6 customers a month and $48,000 in revenue. After a monthly cost of $2,500, that leaves $45,500.
Enquiries a month
60
Customers a month
6
Revenue a month
$48,000
Revenue minus cost
$45,500
Return multiple
19.2x
Revenue for every $1 spent.
Visits needed to break even
209
Visits a month to cover the cost.

How it is worked out

  • Enquiries = visits × enquiry rate.
  • Customers = enquiries × close rate.
  • Revenue = customers × average customer value.
  • Revenue minus cost = revenue − monthly cost.
  • Return multiple = revenue ÷ monthly cost.
  • Visits to break even = monthly cost ÷ revenue per visit. Revenue per visit = enquiry rate × close rate × customer value.
Size this for my company (free proposal)

What is hidden inside a prompt like 'which blockchain is best for private voting in a DAO'?

Three layers are hidden inside it, and good pages separate them. The prompt says 'blockchain', but the public docs we read on 20 October 2026 describe a voting tool, a cryptographic layer and a chain. Each states different strengths and limits. We do not rank them here, because the point is how each project can explain itself. Chalk Labs' own view, not research: pick the layer first (tool such as Snapshot, then privacy method such as MACI or shielded voting), and choose the chain last, based on where your members already hold tokens.

MACI's documentation calls itself infrastructure for private on-chain voting on Ethereum, built because votes on a public chain make bribery easy: voters can prove how they voted 2. It says no one except a trusted coordinator should be able to decrypt a vote, and that it provides collusion resistance only if the coordinator is honest 2. Snapshot's settings guide says its shielded voting offers 'partial privacy': choices stay private during the voting period and are revealed when the proposal closes 3. Shutter's page says the same: votes stay hidden while voting is open, and Snapshot decrypts them once voting ends 5.

Notice what those pages do. They say who must be trusted, when privacy ends and what the tool does not promise. An AI assistant can quote that wording safely, and a reader can check it. Vague wording such as 'fully private' gives neither.

What the prompt mixes together, and what a project page should state (public docs read on 20 October 2026)
LayerWhat public docs sayWhat your page should state
Voting toolSnapshot's docs describe a voting platform for DAOs with gasless voting 20, and shielded voting that keeps choices private until the proposal closes 3What is hidden, until when, and who can see results first
CryptographyMACI's docs describe encryption and zero-knowledge proofs, with a coordinator who must be trusted for collusion resistance 2Who holds keys, what happens if that party fails, and the stated threat model
ChainNeither source above ranks chains, so any ranking an AI gives comes from elsewhereWhich networks you support, with dates, fees measured by a stated method, and known limits

What do AI engines look at when a crypto project is in the answer?

They lean on official pages for facts and on other sites for comparisons. In Ahrefs' test of 20,000 US prompts, official sites took 63.4% of ChatGPT citations for factual questions but 46.4% for comparison questions 6. Profound's study of 11.84 billion citations found brand-owned sites supplied roughly 57% overall, with ChatGPT drawing 47% from them and 30% from earned media 7. Profound gives no crypto-specific split.

Research on what makes content quotable is limited. The Princeton-led GEO paper found its best methods (adding quotations, statistics and source citations) improved visibility by up to 41% on one measure and 28% on another, and similar methods helped on Perplexity.ai. Keyword stuffing did not help. ChatGPT was not tested 9. Google's guide asks for non-commodity content, meaning unique expert or experienced takes that go beyond common knowledge 8. For a protocol, that means measured results, audit scope and honest trade-offs.

Access is the quiet failure. As of Vercel's late-2024 study, none of the OpenAI, Anthropic or Perplexity crawlers it measured ran JavaScript, so docs rendered only in the browser may look empty to them 10. OpenAI recommends allowing OAI-SearchBot in robots.txt to appear in ChatGPT search 11.

What should a Web3 project publish first?

Publish docs that a stranger and an AI assistant can both act on, starting with the facts people get wrong. That means one page per product with the networks supported, contract addresses by network, audit reports with scope and date, a plain-language threat model and a changelog. Put the limits next to the claims, not in a footnote.

Then add comparison and decision pages. A page such as 'private voting options for DAOs' can cover several tools fairly, including ones that are not yours. It can use the three-layer split above and say when each choice fits. See how to write comparison pages that get cited and how to build a glossary that AI engines cite.

A machine-readable docs index costs little. Snapshot's docs publish an llms.txt file that lists about 66 pages 4. But Google says Google Search ignores llms.txt 8, and SE Ranking found no correlation between having one and AI citations across nearly 300,000 domains 18. Treat it as a tidy extra, not a ranking lever, and spend your time on the pages themselves.

A limits box for a privacy feature (placeholders)
  • What is private: individual votes, while voting is open.
  • What is not private: totals, the final result, and the voter's wallet address.
  • Who must be trusted: [the coordinator or key holders], who can [decrypt votes if they collude].
  • When privacy ends: [at proposal close / never / after N days].
  • What we do not claim: protection from [bribery off-chain / compromised wallets].
  • Audited by: [firm], scope [contracts audited], dated [date], report at [link].
  • Last reviewed: [date]. Report errors to [email].

How do PR and paid placements fit into AI visibility?

Independent coverage helps AI judge a project, but pay-to-play placements carry risk. ChatGPT drew about 30% of its citations from earned media in Profound's data 7, and Google says chasing inauthentic mentions is not as helpful as it seems 8. So aim for real news: launches, audits, research and named expert quotes.

Google's site reputation policy targets third-party content hosted on a site mainly to borrow that site's ranking signals 12. It lists wire and press-release services, syndicated news, editorial columns and third-party advertorials meant to reach readers as not inconsistent with the policy 12. Google says no single factor decides the outcome, so a disclosure label alone is not a safe harbour 12.

Disclosure is also a legal matter. A 2017 SEC statement says endorsements of investments, including coin offerings that may be securities, may be unlawful if they do not disclose the nature, source and amount of any compensation 13. Label sponsored posts, keep a record of what you paid for, and never present paid praise as independent review.

What rules apply to crypto marketing and claims?

Rules vary by country, and your website reaches all of them. The UK's Financial Conduct Authority says that since 8 October 2023 its remit includes qualifying cryptoassets, and that its promotion rules apply to all firms marketing cryptoassets to UK consumers, whether or not the firm is based overseas 14. It lists four lawful routes, such as promotion by an authorised person or approval by one 14.

In the EU, the European Securities and Markets Authority says MiCA disclosures take the form of a white paper, and its register notes that white papers are not reviewed or approved by any competent authority 15. The 2017 SEC statement concerns paid promotion of coin offerings and other investments that may be securities 13. These are examples, not a full list.

A practical rule helps: if a sentence could change what someone does with money or keys, give its source, date and limit in the same paragraph. Leave out price predictions and return claims. Have counsel approve pages that say who can use a product and where.

Examples of rules that can touch crypto web pages (not legal advice)
MarketRule or sourceWhat it means for your pages
United KingdomFCA regime for qualifying cryptoassets, from 8 October 2023 14Check who can lawfully communicate a promotion to UK consumers, even from overseas
United StatesSEC 2017 statement on paid promotion of coin offerings and other investments that may be securities 13Disclose the nature, source and amount of any payment behind an endorsement
European UnionMiCA white paper regime, per ESMA 15Keep your white paper page current, and do not imply regulators approved it
EverywhereGoogle site reputation policy 12Do not host or buy third-party pages mainly to borrow a site's ranking signals

How do you measure AI visibility for a crypto project?

Measure whether you are named and whether the answer is correct. Run five core prompts on two AI engines, three runs each, and score each as named, cited or absent. Add an accuracy column for chains supported, contract addresses and audit status, because a wrong fact here can cost users money.

Use free reports where they exist. Google's Search Console generative AI report shows impressions in AI Overviews and AI Mode by page, country, date and device, and its help page mentions no click data 17. Add a question to your community onboarding and partner forms: 'Where did you first hear about us?' with a free-text answer, since many discovery paths in Web3 run through chats and forums that analytics cannot see.

If an engine gives a wrong address or claims support for a chain you do not support, correct the source first, then your docs, then rerun the prompt. Our guide to fixing wrong information AI says about your brand lists the steps, and the Query Fan-Out Explorer helps you see what searches a prompt may trigger.

What have Chalk Labs and Rahil Jain done in Crypto and Web3, and what have they not?

Web3 and AI is Chalk Labs' home market, and Rahil Jain is a co-founder of Chalk Labs, a Web3 and AI growth agency. Its published client work in this field is PR: press placements for Web3 and AI companies. Its AI SEO case studies are a local shop and a co-founder's personal site, and none reports AI-citation counts. We name no clients on this page.

The public record is narrower than that sounds. Of seven case files on Chalk Labs' case-studies page, five describe press placements for Web3 and AI companies and none reports AI-citation counts 16. Two describe search work: a Bengaluru trophy shop and shilikajain.com, the personal site of Chalk Labs co-founder Shilika Jain 16. Chalk Labs says that site ranked first for 'fractional PR manager crypto' within a month of launch, but the page shows no search volume or ranking history 16.

So we have not published AI-citation results for Web3 clients. Rahil Jain runs Chalk Labs' AI SEO programme, and Chalk Labs sells this service, so we have a commercial interest in this topic.

Your first 90 days

Days 1-30

  • Run ten buyer and builder prompts on two engines, three runs each, and score accuracy.
  • Check that docs render without JavaScript and that robots.txt and your CDN allow AI search bots.
  • List every public claim about security, privacy and returns, and have counsel review each.

Days 31-60

  • Publish a limits box and a threat model for each core feature.
  • Publish a networks and addresses page and an audits page with scope and dates.
  • Add a free-text 'where did you hear about us?' question to community and partner forms.

Days 61-90

  • Publish one fair decision page that compares several approaches, including others' tools.
  • Pitch two pieces of real news, such as an audit result or research, with sponsorship disclosed.
  • Rerun the prompt log, compare it with day one and fix any wrong answers at their source.

The first pages to publish

  1. Networks supported and contract addresses by network. Wrong addresses are the costliest 'almost right' answer, and facts come from official sites 6.
  2. Threat model and limits for each feature. MACI's docs show how stating who must be trusted makes wording quotable 2.
  3. Audit reports with firm, scope and date. It gives AI and users a checkable trust signal.
  4. A fair decision page for your category. Comparison questions lean on outside sources, so be the neutral source 6.
  5. Docs changelog. It shows which version a statement applies to.
  6. Governance and team page. It names who stands behind the project.
  7. Markets served and rules followed. It states where you promote, since UK rules reach overseas firms 14.
  8. Glossary of your protocol's terms. Short definitions are easy to quote and reduce misreadings.

Trust signals and risks

What AI looks for

  • Docs with exact versions, networks, contract addresses and dates.
  • Audit reports that state scope, firm and date.
  • A threat model that names who must be trusted and when privacy ends 2.
  • Named team members with real profiles and roles.
  • Independent coverage with disclosed sponsorship.
  • A changelog and a public error-report address.

What can go wrong

  • Price predictions, return claims or unproven security claims.
  • Undisclosed paid promotion, which the SEC says may be unlawful 13.
  • Third-party pages hosted mainly to borrow a site's rankings 12.
  • Marketing to UK consumers without a lawful route, even from overseas 14.
  • Docs that bots cannot read, or wrong facts that AI keeps quoting 10.

Why Rahil Jain for Crypto and Web3?

Direct experience

Chalk Labs has done work in this field and can show it.

Web3 and AI is Chalk Labs' home market. Its published client work in this field is PR: press placements for Web3 and AI companies. Its AI SEO case studies are a local shop and a co-founder's personal site, and none reports AI-citation counts. We name no clients here. Chalk Labs sells this service, so weigh our view accordingly.

Rahil Jain

Rahil Jain is an AI SEO (answer engine and generative engine optimisation) consultant and co-founder of Chalk Labs.

LinkedIn · Method and credentials

Results we can show

These are Chalk Labs' own search results. They are not a forecast for your site: results depend on your market, competition and starting point.

27.3k
Google clicks in 16 months for a local trophy shop in Bengaluru (Meru Trophy)
2.3M
search impressions over the same period, at an average position of 6.9
#1
ranking for “fractional PR manager crypto” within a month of launch (shilikajain.com, a low-competition niche phrase)

How to read these numbers

Chalk Labs reports them from Google Search Console data. We cannot publish the raw console. The shilikajain.com site belongs to Shilika Jain, a Chalk Labs co-founder, so treat it as a demonstration on our own network and not as an independent client result. Meru Trophy is a client of Chalk Labs.

See the case studies

Frequently asked questions

Does ChatGPT recommend crypto and Web3 projects?

It can name them, but nobody can promise it will name yours. Official sites supplied 63.4% of ChatGPT citations on factual questions but only 46.4% on comparisons 6. So docs carry the facts and independent coverage carries the comparisons. Answers vary by run, and we do not guarantee placement.

Do I need an llms.txt file for my protocol's docs?

It is optional. Google says Google Search ignores llms.txt 8, and SE Ranking found no correlation between having one and AI citations across nearly 300,000 domains 18. Put your effort into clear pages first.

Can paid crypto media placements help AI visibility?

Real coverage can help, but pay-to-play hosting can backfire. Google's site reputation policy targets third-party content hosted mainly to borrow a site's rankings, though wire services, syndicated news, columns and readable advertorials are listed as not inconsistent 12. Disclose sponsorship, as the SEC expects for paid promotion 13.

What should a token project avoid saying on its website?

Avoid price predictions, return claims and unproven security claims. UK rules reach firms marketing cryptoassets to UK consumers even from overseas 14, and the SEC says undisclosed paid endorsements may be unlawful 13. Have counsel review each claim. This is general information, not legal advice.

Do developers trust AI answers about Web3 code?

Only partly. In Stack Overflow's 2025 survey, 45.7% distrusted the accuracy of AI output and 32.7% trusted it, and 66% of respondents named answers that are almost right but not quite as the most common frustration 1. Publish exact versions, addresses and examples so answers stay checkable.

How do I check what AI says about my protocol?

Run five core prompts monthly on two engines, three runs each. Score each answer as named, cited or absent, and add an accuracy check for chains, addresses and audits. Fix wrong facts at their source first, then on your own docs, and rerun the prompt to confirm.

Sources

  1. 2025 Stack Overflow Developer Survey: AI · Stack OverflowOpened October 20, 2026. Supports: 84% of respondents use or plan to use AI tools in development (47.1% daily); 45.7% distrust AI output accuracy against 32.7% who trust it; 66% name AI solutions that are almost right, but not quite as their top frustration.
  2. MACI documentation: Introduction · Privacy and Scaling Explorations (Ethereum Foundation)Opened October 20, 2026. Supports: MACI is described as infrastructure for private on-chain voting; public chain votes make bribery easy because voters can prove their choice; no one except a trusted coordinator should be able to decrypt a vote; collusion resistance depends on an honest coordinator; it presumes each member controls a unique Ethereum private key.
  3. Snapshot documentation: Space settings (shielded voting) · Snapshot LabsOpened October 20, 2026. Supports: Space settings page: shielded voting offers partial privacy, uses threshold encryption so choices are private during the voting period and revealed when the proposal closes, and is powered by Shutter.
  4. Snapshot documentation index (llms.txt) · Snapshot LabsOpened October 20, 2026. Supports: The docs site publishes an llms.txt file, a Markdown index of about 66 links, hosted on Mintlify.
  5. Shielded voting · Shutter NetworkOpened October 20, 2026. Supports: Votes stay hidden during the voting period, and Snapshot decrypts all votes and reveals the results once voting ends.
  6. How ChatGPT's query fan-out works (Ahrefs study of 66.6 million responses) · AhrefsOpened October 20, 2026. Supports: Article dated 12 October 2026: in 20,000 US prompts official sites took 63.4% of ChatGPT citations for fact questions and 46.4% for comparison questions.
  7. Where do AI citations come from? · ProfoundOpened October 20, 2026. Supports: Article dated 30 July 2026 on 11.84 billion citations (16 April to 16 July 2026): brand-owned sites about 57% of citations overall; ChatGPT 47% brand-owned and 30% earned; no crypto-specific figure is stated. Profound sells AI visibility tracking.
  8. Optimizing your website for generative AI features on Google Search · Google Search CentralOpened October 20, 2026. Supports: Last updated 10 July 2026: Google Search ignores llms.txt; non-commodity content offers unique expert or experienced takes; seeking inauthentic mentions is not as helpful as it seems.
  9. GEO: Generative Engine Optimization · arXiv (Aggarwal et al.; accepted at ACM SIGKDD 2024)Opened October 20, 2026. Supports: Best methods (adding quotations, statistics and source citations) improved visibility by up to 41% on Position-Adjusted Word Count and 28% on Subjective Impression in the authors' test engine; similar methods helped on Perplexity.ai; keyword stuffing did not help; efficacy varies across domains; ChatGPT not tested.
  10. The rise of the AI crawler · Vercel (with MERJ)Opened October 20, 2026. Supports: Article dated 17 December 2024: none of the major AI crawlers measured from OpenAI, Anthropic, Meta, ByteDance or Perplexity rendered JavaScript; Gemini and Apple's crawler did. Behaviour may have changed since.
  11. Overview of OpenAI Crawlers (bots) · OpenAIOpened October 20, 2026. Supports: OAI-SearchBot surfaces sites in ChatGPT search and OpenAI recommends allowing it in robots.txt; sites that opt out will not be shown in ChatGPT search answers.
  12. Spam policies for Google web search (site reputation policy) · Google Search CentralOpened October 20, 2026. Supports: Last updated 28 August 2026: the policy covers third-party content published mainly to borrow a host's ranking signals; wire and press-release services, syndicated news, editorial columns and third-party advertorials meant to reach readers are listed as not inconsistent; the text says no single factor is necessary or sufficient and lists disclosure and editorial responsibility as supporting factors.
  13. Statement on Potentially Unlawful Promotion of Initial Coin Offerings and Other Investments by Celebrities and Others · US Securities and Exchange CommissionOpened October 20, 2026. Supports: Dated 1 November 2017: coins sold in ICOs may be securities; endorsements may be unlawful if they do not disclose the nature, source and amount of any compensation.
  14. Cryptoassets (financial promotions) · UK Financial Conduct AuthorityOpened October 20, 2026. Supports: Last updated 30 June 2026: qualifying cryptoassets came into the FCA's remit on 8 October 2023; four lawful routes for promoting cryptoassets to consumers; the regime applies to all firms marketing cryptoassets to UK consumers, whether or not based overseas.
  15. Markets in Crypto-Assets Regulation (MiCA) · European Securities and Markets AuthorityOpened October 20, 2026. Supports: MiCA disclosures take the form of a white paper; the interim register states that white papers have not been reviewed or approved by any competent authority in any Member State.
  16. Chalk Labs case studies · Chalk Labs (self-published)Opened October 20, 2026. Supports: Seven case files: five describe press and PR placements for Web3 and AI companies, two describe SEO, AEO and GEO work (Meru Trophy and shilikajain.com); none reports AI-citation counts; shilikajain.com is reported to rank first for 'fractional PR manager crypto' within one month of launch, with no search volume or ranking history shown.
  17. Generative AI performance report (Search Console Help) · Google Search Console HelpOpened October 20, 2026. Supports: The report shows impressions in AI Overviews and AI Mode by page, country, date and device; the help page mentions no click data; rolled out to all websites worldwide as of 31 August 2026.
  18. LLMs.txt study · SE RankingOpened October 20, 2026. Supports: Article dated 15 October 2026: nearly 300,000 domains, 10.13% with an llms.txt file, and no correlation between having one and AI citation frequency.
  19. AI features and your website · Google Search CentralOpened October 20, 2026. Supports: A page must be indexed and eligible to show with a snippet to be a supporting link in AI Overviews or AI Mode; no additional technical requirements or special schema.org markup are needed.
  20. Snapshot documentation home · Snapshot LabsOpened October 20, 2026. Supports: Snapshot describes itself as a voting platform that lets DAOs, DeFi protocols and NFT communities vote easily and without gas fees.
How this page was made. Built on Rahil Jain's AI SEO method and drafted with AI research assistance. Every statistic links to the page we read it on (opened October 20, 2026 or nearby). Published October 20, 2026; last updated October 20, 2026. Chalk Labs sells AI SEO services, so we have a commercial interest in this topic. Spotted an error? Tell us and we will fix it.