Lawyers, Doctors, Experts: AI Visibility and YMYL Content

A patient types "left chest pain" into ChatGPT. An employee asks Perplexity "how to contest an unfair dismissal." A taxpayer asks Gemini about "tax optimization for a real estate LLC." In all three cases, the quality of the answer can have direct consequences on the person's health, freedom, or finances.

Google calls these topics YMYL: Your Money or Your Life. Content whose inaccuracy can cause real harm. And this framework, designed for traditional search results, has become even more critical with AI engines.

AI engines don't just list links. They synthesize, rephrase, and sometimes hallucinate. In a YMYL domain, a hallucination isn't an inconvenience — it's a danger. That's precisely why AI engines apply reinforced credibility filters on these topics.

Good news for regulated professions: lawyers, doctors, accountants, architects, and notaries have a considerable natural advantage. Their degrees, professional registrations, and ethical frameworks are exactly what Google expects from a YMYL source. What remains is turning them into content AI engines can cite.

YMYL: why it's the most demanding terrain for AI

What exactly is YMYL?

YMYL (Your Money or Your Life) is a classification defined by Google in its Search Quality Rater Guidelines. It designates topics whose content can significantly affect:

• Health: symptoms, diagnoses, treatments, medications, mental health
• Financial security: investments, taxes, insurance, loans, retirement (see our article on what ChatGPT recommends on banking and insurance)
• Legal security: rights, procedures, contracts, disputes, criminal law
• Physical safety: safety advice, emergency situations
• Social well-being: civic information, elections, public services

For these topics, Google requires a higher level of evidence. Quality Raters are instructed to systematically verify the author's credentials, the source's reliability, and compliance with professional consensus.

Google says so explicitly: its systems give even more weight to content that aligns with strong E-E-A-T for topics that could significantly impact the health, financial stability, or safety of people. This is a Google ranking principle, which matters for AI engines insofar as they rely on search engine results. The same page specifies that E-E-A-T is not itself a ranking factor. And the Detekia calibration study (317 pages from large French sites) did not confirm it as a page-level citation factor: a named author did not distinguish cited pages from other pages on the same site. On YMYL topics, credibility is expected; what gets a page cited on top of that is how it answers.

Why AI engines are even stricter on YMYL

AI engines have an additional problem compared to Google: they don't just recommend links, they generate answers. When ChatGPT synthesizes medical, legal, or financial information, it implicitly stakes its credibility. An error isn't a bad ranking in a list of results. It's false information presented as fact.

This is why RAG systems (Retrieval-Augmented Generation) apply reinforced filters on YMYL queries:

• More restrictive source selection. AI engines favor institutional sources, identified professional sites, and reference publications. A law firm's website with a named author and bar registration inspires more trust than an anonymous blog about employment law, with readers and in Google's ranking alike.
• Enhanced cross-verification. On a medical query, AI engines triangulate more: they look for convergence between multiple reliable sources before formulating an answer. Isolated information, even if correct, is less likely to be cited.
• Systematic disclaimers. ChatGPT, Gemini, and Perplexity add disclaimers on YMYL topics ("consult a healthcare professional," "this does not constitute legal advice"). But they still cite sources, and those sources are the ones that pass the credibility filter.

→E-E-A-T and AI: what actually gets you cited

The natural advantage of regulated professions

Regulated professions (lawyers, doctors, certified accountants, architects, notaries, pharmacists) possess an asset that most content creators lack: verifiable proof of expertise by default.

E-E-A-T signals built into regulated professions

Every regulated professional naturally has:

• A state-recognized degree. Medical doctorate, bar exam, CPA certification. These are verifiable credentials.
• Professional registration. Medical board, bar association, accounting board. These registrations appear in public directories, which your clients and search engines can both consult.
• An ethical framework. The obligation of competence, continuing education, and professional liability constitutes a Trustworthiness signal in Google's sense.
• Documentable field experience. Years of practice, specializations, cases handled. These are the Experience proofs that the E-E-A-T framework requires.

The problem is that most professionals don't make these signals visible online. A lawyer with 20 years at the bar and a specialization in business law may have a website that displays none of this information in a structured way. For AI engines, it's as if these credentials don't exist.

The YMYL paradox: the most qualified are often the least visible

The professionals most qualified to discuss YMYL topics are often the least present online. Several reasons explain this paradox:

• Ethical constraints limit communication (regulated advertising for lawyers and doctors)
• Word-of-mouth remains the dominant acquisition channel for many firms and practitioners
• Lack of time and digital skills slows web investment
• Distrust of "digital marketing" in professions built on trust relationships

Result: AI engines cite less qualified sources (general information sites, forums, blogs), which publish clear answers to the questions people ask, while real experts publish little. It's a loss for users and a missed opportunity for professionals.

How AI engines evaluate YMYL content credibility

Understanding the AI selection mechanism on YMYL topics allows you to identify precise action levers. Here's how the trust chain works.

Step 1: filtering by the underlying search engine

ChatGPT no longer depends on Bing alone: it combines its own index, fed by its OAI-SearchBot crawler, with several external providers, including Microsoft and Google (Peec AI, 2026). Gemini relies on Google, Perplexity on a hybrid index. These engines already apply a YMYL filter: for a query like "stroke symptoms," a search engine will favor recognized medical sources (NHS, Mayo Clinic, hospital sites, practitioner sites).

If your site doesn't pass this first filter, it will never reach the LLM. That's why SEO fundamentals remain important in GEO.

Step 2: selection by the LLM

Once results are retrieved by the search engine, the LLM applies its own selection layer. On YMYL topics, this selection is stricter:

• Author identification. A named author with verifiable credentials ("Dr. Sarah Johnson, cardiologist, Mayo Clinic") reassures readers and meets what Google expects on YMYL topics. But in the Detekia calibration study, a named author did not separate cited pages from the rest: it is a credibility baseline, not a citation lever on its own.
• Consensus consistency. The LLM checks whether the information is consistent with the medical, legal, or financial consensus it knows. Content that contradicts consensus without justification will be discarded.
• Structure and citability. Content that places a short answer under each heading, phrases its headings as questions, and uses tables is easier to extract and cite: these are the traits that distinguished cited pages in the study, along with a recent update date and complete content. JSON-LD schemas (Person, MedicalEntity, LegalService) describe the page without being decisive.

Step 3: response formulation with disclaimers

Even when an AI cites your YMYL content, it adds disclaimers. But the citation is there. And the user who reads "according to Attorney Smith, a member of the New York Bar, the statute of limitations for wrongful termination claims is typically one year (source: smithlaw.com)" has access to sourced, verifiable information.

This citation is precisely what drives qualified traffic to your site. And it's the most valuable traffic: people in real need, looking for a competent professional.

Specific risks: hallucinations and erroneous recommendations

YMYL domains are where AI hallucinations have the most serious consequences. Understanding these risks reinforces the value of your presence as a reliable source.

Medical hallucinations

AI can invent drug interactions, suggest incorrect dosages, or minimize serious symptoms. Several studies published since 2023 show that LLMs still provide partially incorrect medical information on a significant share of complex questions. When a practitioner publishes structured, sourced content, they help reduce this rate by providing AI engines with reliable data to cite.

Legal hallucinations

AI engines regularly confuse jurisdictions, cite repealed statutes, or invent case law. The widely publicized case of New York attorney Steven Schwartz (2023), whose brief contained case citations fabricated by ChatGPT, illustrates the risks. Legal content published by lawyers with precise references (statutes, dated case law, official sources) is the best protection against these errors.

Financial hallucinations

Outdated interest rates, incorrect tax thresholds, poorly explained investment mechanisms. Financial content is particularly sensitive because regulations change frequently. An accountant who maintains up-to-date content with visible modification dates provides a freshness signal that AI engines strongly value.

For every type of professional, an online presence isn't a luxury. It's a service to users who, without your reliable content, receive potentially inaccurate AI answers.

These inaccuracies can also target you directly: an AI that attributes a specialty you don't practice to your firm, or that consistently points to a peer in your field. As in e-commerce, where AI engines arbitrate between competing brands (see how AI engines choose the products they recommend), they cite a few names and ignore others. If peers are cited instead of you, see what to do when ChatGPT cites your competitors.

5 concrete actions for YMYL professionals

Here are the 5 highest-impact actions for a lawyer, doctor, accountant, or any regulated professional who wants to be cited by AI engines on their areas of expertise.

1. Create a complete author page with degrees and specializations

The author page is the foundation of your credibility with readers, and Google expects it on YMYL topics. It should contain:

• Full identity: name, professional title (Esq., Dr., CPA, etc.)
• Degrees and training: university, year, specializations, certifications
• Professional registration: registration number, bar or board affiliation
• Experience: years of practice, specialization areas, career path
• Publications: articles in specialized journals, presentations, conferences
• Professional photo: an identifiable face strengthens trust

Every piece of content published on your site should link to this author page. It's what allows readers, and search engines, to verify the author's credentials.

2. Implement the Person Schema with medical or legal credentials

The JSON-LD Person Schema describes an author's identity and qualifications in a standardized format. Useful, but not decisive: for its AI answers, Google states that no special markup is required. If you implement it, go beyond the basic schema. Use Physician for doctors with medicalSpecialty, qualifications, and memberOf linking to the medical board. For lawyers, use LegalService alongside the Person schema, with bar affiliation and practice areas.

→Schema.org and AI: practical guide for LLMs

3. Publish structured medical, legal, or financial FAQs

FAQs are the most directly citable format for AI engines. A precise question with a concise, sourced answer is exactly what RAG systems look to extract.

Rules for effective YMYL FAQs:

• Phrase questions as your patients or clients ask them. Not "What are the modalities of the termination procedure for personal reasons?" but "Can my employer fire me without cause?"
• Answer in 2-3 factual sentences before expanding. The first sentence should be standalone and citable.
• Cite your sources. Statutes, clinical guidelines, case law. AI engines cite content that itself cites sources.
• Add the FAQPage schema, identical to the visible FAQ. It describes the question/answer pairs; the visible FAQ is what matters first.
• Date each FAQ. "Updated May 15, 2026" is a decisive freshness signal for AI engines on regulatory topics.

A law firm that publishes 20 structured FAQs on employment law, with FAQPage schema and statutory references, becomes a go-to source for AI engines on these queries.

4. Be present in professional directories and reference databases

Professional directories play a key role in AI triangulation. When an AI engine verifies an author's credibility, it looks for consistent mentions across multiple independent sources.

Priority directories by profession:

• Doctors: state medical board directories, Healthgrades, Zocdoc, hospital directories
• Lawyers: state bar directories, Martindale-Hubbell, Avvo, Chambers, Legal 500
• Accountants: AICPA directory, state CPA society directories
• Architects: AIA directory, state licensing board directories

Make sure your name, specialization, and website URL are consistent across all directories. Inconsistencies (different name, different address, different specialization) weaken the trust signal.

5. Publish and get cited in specialized press

Publications in professional journals and citations in specialized press are the most powerful Authoritativeness signals for AI engines.

Concrete actions:

• Publish in your profession's journals. Law reviews, medical journals, accounting publications. Even a case commentary counts.
• Respond to journalists. Platforms like HARO allow you to position yourself as an expert source. Every citation in a press article strengthens your authority in the eyes of AI engines.
• Participate in conferences and webinars. Replays and published summaries are expertise signals that AI engines capture.
• Write op-eds in online media. Major publications regularly feature expert opinions. A single publication in a reference media outlet can transform how AI engines perceive your authority.

The goal is to create a network of mentions confirming your expertise. AI engines use triangulation: when your name appears as an expert on your site, in a professional directory, in a specialized journal, and in a press article, the signal is unambiguous.

→GEO score: how to measure your site's AI visibility

Illustrative example: a law firm, step by step

This example is fictional. It shows how the actions above fit together for a typical firm; it does not describe the results of a real client.

Starting point

A law firm specializing in employment law, with three partners. The website has five pages (home, practice areas, team, contact, legal notice): no editorial content, no JSON-LD schema, a team page with first names and photos but no degrees or bar registration.

When someone asks ChatGPT or Perplexity for an "employment lawyer" in the firm's city, a site like this is unlikely to appear: AI engines tend to cite directories and general information sites instead.

Actions to implement

1. Individual author pages for each partner, with degrees, bar registration, specializations, publications, and enriched Person schema
2. Structured FAQs on employment law, with FAQPage schema, statutory references, and update dates
3. LegalService schema on the homepage with address, bar affiliation, practice areas, hours
4. Directory harmonization: up-to-date profiles on the bar's website, LinkedIn, Google Business Profile and legal directories
5. Op-eds in specialized media with a full bio and a link to the site

How to measure the effect

Before starting, write down the AI answers to a dozen of your clients' questions ("deadline to contest a dismissal," "unfair dismissal compensation"). Ask the same questions again after a few weeks, on ChatGPT, Perplexity and Gemini, and compare: is your firm cited, in which position, and which sources appear next to you? Results depend on your local competition and on your site's starting authority: no timeline or gain is guaranteed.

How Detekia helps YMYL professionals

Law firms, healthcare practitioners, and accountants don't have time to become GEO specialists. That's precisely Detekia's role.

Our free GEO score analyzes your site on 6 AI citability criteria and highlights the gaps that weigh most heavily on YMYL sites:

• Readable by AI (prerequisite): do OAI-SearchBot, Claude-SearchBot and PerplexityBot have access to your content? Without it, the score is capped at 40
• Answer structure: a short answer under each heading, headings phrased like your clients' questions
• Evidence & data: numbers, linked external sources and attributions, essential on medical, legal or financial topics
• Freshness: publication and modification dates, update signals
• Entity clarity and external signals: a clearly identified firm name, a Wikidata entry, client reviews
• Citation test: 5 queries put to one of OpenAI's models to check whether your site gets cited

You get a free score out of 100, a breakdown of the 6 criteria and prioritized recommendations. To measure your actual presence in ChatGPT, Gemini, Claude and Perplexity, an audit of what AI engines say about your practice takes it from there. YMYL professionals often start with a low score (due to a lack of editorial content) but progress quickly: their credentials are already there, they just need to be made visible to AI engines.

Conclusion: credibility is your best GEO asset

Regulated professions have a structural advantage in the race for AI visibility. Degrees, professional registrations, publications, field experience: everything that constitutes your professional legitimacy is what Google expects from a YMYL source. To get cited, put it to work in clear answers: a short answer under question headings, sources, an update date.

The challenge isn't creating credibility. It already exists. The challenge is making it machine-readable: structured author pages, enriched JSON-LD schemas, sourced FAQs, consistent presence in professional directories.

3 actions to start this week:

1. Create or enrich your author page with degrees, bar/board registration, and specializations. Add the corresponding Person schema.
2. Publish 5 structured FAQs on your most frequent client questions, with sources and FAQPage schema.
3. Verify the consistency of your information across the 3 main directories of your profession.

→GEO: the complete guide to getting cited by AI in 2026

→Schema.org and AI: practical guide for LLMs