Guide · Senior-Leader

GEO in Regulated Industries: Health, Finance and Legal

Regulated firms can do GEO, but every page written for AI engines is still marketing and must meet the same rules. The specific risk is context collapse: an AI answer may quote one sentence and drop the disclaimer under it. So in health, finance and legal, write each key claim so it stays accurate when quoted alone, and run GEO content through the same compliance review as any other marketing.

Method: quotes are from EU primary law as published in the Official Journal: Directive 2014/65/EU (MiFID II), Directive 2005/29/EC (UCPD) and Regulation (EU) 2024/1689 (AI Act), read on 31 October 2026 (CET). National rules and sector regulators add more. This isn't legal advice. Opinions are marked "Our view".

On this page

Checked on 31 October 2026. "Our view" marks MentionX opinion. Ledgerline is a fictional company used for illustration.

Why are regulated industries different for GEO?

Three reasons:

  1. Rules on how claims are made. Regulated marketing often has specific rules on balance, risk warnings and identification.
  2. Higher harm. A wrong AI answer about a drug interaction or a pension product can hurt someone.
  3. Context collapse. Engines extract passages. A risk warning three paragraphs below a claim may never make it into the answer.

Our view: the third point changes how you write, more than any rule does.

Which rules apply to all sectors?

The Unfair Commercial Practices Directive applies to business-to-consumer marketing generally. A practice is misleading if it "deceives or is likely to deceive the average consumer, even if the information is factually correct", including about "the main characteristics of the product, such as its... benefits, risks... or the results to be expected from its use" (Directive 2005/29/EC, Art. 6(1)(b)).

Omissions count too. A practice is misleading if it "omits material information that the average consumer needs, according to the context, to take an informed transactional decision" (Directive 2005/29/EC, Art. 7(1)).

Annex I bans undisclosed paid editorial content ("advertorial", point 11) and "falsely representing oneself as a consumer" (point 22) in all cases (Directive 2005/29/EC, Annex I). See our GEO ethics guide.

The AI Act's disclosure duties also apply. Under Article 50(4), deployers must disclose AI-generated text published to inform the public on matters of public interest, unless it has had human review or editorial control and someone holds editorial responsibility (EU AI Act, Art. 50(4)). Health and finance explainers can fall in that category. See the EU AI Act for marketers.

What does finance regulation require of GEO content?

For EU investment firms, MiFID II Article 24(3) states: "All information, including marketing communications, addressed by the investment firm to clients or potential clients shall be fair, clear and not misleading. Marketing communications shall be clearly identifiable as such" (Directive 2014/65/EU, Art. 24(3)).

Article 24(1) sets the wider standard: an investment firm must "act honestly, fairly and professionally in accordance with the best interests of its clients" (Directive 2014/65/EU, Art. 24(1)).

Two parts of Article 24(3) matter for GEO:

RequirementGEO implication (Our view)
"fair, clear and not misleading"A quotable sentence about returns must carry its own balance. Don't put the upside in one paragraph and the risk in another
"clearly identifiable as such"Educational hubs that are really product marketing should say so. Don't dress product pages as neutral guides

Fictional example: instead of "Our fund returned 9% last year." followed by a risk section further down, write "Our fund returned 9% in 2025; past performance does not predict future returns and you could lose money." If an engine quotes that sentence, the balance travels with it.

What about health content?

Health marketing in the EU is governed by sector rules (for example on medicines and medical devices) and national law, which we haven't quoted here. Check with your regulatory team. The UCPD principles above still apply to consumer health marketing.

Our view, for health GEO:

  • Name the reviewer. Show a qualified reviewer and review date on clinical content. That's good practice for readers and engines alike.
  • Separate education from promotion. Keep disease-awareness content distinct from product pages, and label each.
  • Write self-contained safety statements. Contraindications and "talk to your doctor" guidance belong next to the claim, not only in a footer.
  • Don't optimise for prompts you can't answer safely. If the honest answer is "see a clinician", say that clearly.

Rules on lawyers' advertising are set nationally and by bar associations, so we haven't quoted them here. The UCPD applies to consumer-facing legal marketing.

Our view:

  • Say where you're qualified. AI answers lose geography. "Our employment lawyers in Germany advise on…" holds its meaning out of context better than "We advise on…".
  • Separate general information from advice. Mark guides as general information, and say so inside the key paragraphs, not only in the footer.
  • Avoid outcome claims. Statements about results can be misleading under Art. 6 if they imply future success.

How should a regulated GEO workflow run?

  1. Map claims. List every product, performance or clinical claim on pages you want cited.
  2. Rewrite for extraction. Make each claim self-contained, with its qualifier in the same sentence.
  3. Add compliance review. GEO content goes through the existing marketing sign-off. Record the reviewer and date.
  4. Disclose AI assistance where Art. 50(4) applies, or make sure the human-review exemption is met and documented.
  5. Monitor AI answers. Track what engines say about your products. When an answer is wrong, fix your source pages first; our crisis playbook covers escalation.
  6. Decide crawler access deliberately. See blocking or allowing AI crawlers.

What are the stakes if AI disclosure is missed?

Under the AI Act, breaching the Article 50 transparency obligations can bring fines "up to EUR 15 000 000 or, if the offender is an undertaking, up to 3 % of its total worldwide annual turnover for the preceding financial year, whichever is higher" (EU AI Act, Art. 99(4)). Sector regulators can add their own sanctions on top. Our view: in regulated firms, the reputational cost of a misleading AI answer traced back to your page is usually the bigger risk.

What should a regulated GEO team avoid?

TacticWhy to avoid it
Paid "best fund" or "best clinic" list placements without labelsUCPD Annex I point 11
Staff posting as patients or customersUCPD Annex I point 22
Claims that only make sense with a footnoteFail "fair, clear and not misleading" once extracted (Our view)
Unreviewed AI explainers on health or moneyAI Act Art. 50(4), plus accuracy risk

From MentionX data: which questions name brands

Method. Read-only aggregate queries over MentionX scans from 22 August to 29 October 2026 (UTC): 2,217 engine runs (2,154 completed) across 18 anonymised projects and six engines.

Our sample isn't split by industry, so we can't report regulated sectors separately. It shows how often brands are named across question types:

IntentBrand mentionedShare of voice
Brand-specific100%64.3
Comparison91.9%44.6
Recommendation29.2%26.8
Best-of26.6%22.7
Problem5.7%7.7

And how often engines cite sources:

EngineBrand mentionedAnswers with no citations
Gemini47.3%50.7%
OpenAI42.9%30.2%
Perplexity31.5%4.1%
AI Mode31.3%12.6%
Copilot26.0%47.0%
DeepSeek24.0%99.0%

Our view: in regulated sectors, a brand named in a recommendation or comparison answer can be read as a claim about suitability. That raises the stakes on accuracy. Where engines cite, check what your cited pages say; where they don't, the facts on your own and third-party pages still shape the answer.

Limits. A convenience sample of mostly B2B software projects, run through engine APIs, which can behave differently from the consumer apps. Perplexity, Copilot and DeepSeek figures come from a single day (26 October 2026), so treat them as a snapshot, not a trend.

Worked example: A fictional finance team adapts answer-first writing

Ledgerline is a fictional accounts-payable tool. Figures are illustrative.

Ledgerline isn't regulated in this way, so this example uses a fictional savings app. Its GEO lead wants pages that open with a direct answer. Compliance points out that the interest rate must appear with its conditions and that past returns can't be shown without the required warnings.

They agree a pattern: the first sentence gives the answer and its key condition together, such as the rate and the minimum balance. The required warning sits directly below, not in a footer. Every figure links to its source page and has a review date. Compliance signs off a template once, which speeds up later pages.

How do you keep qualifications next to claims?

AI answers often quote a single sentence. If the claim is in one sentence and its condition is three paragraphs away, the answer may carry the claim without the condition. Write claims and their key qualifications in the same sentence, or at least the same short paragraph. This is good practice in any sector and especially important in regulated ones.

Who should review GEO content?

The same people who review other public content: compliance, legal or clinical reviewers, depending on the sector. Add GEO work to their existing workflow rather than creating a separate one. Give reviewers a short explanation of why pages are written answer-first, so they understand the format.

How do you handle third-party descriptions?

Comparison sites and forums may describe your products in ways you couldn't say yourself. You can ask for factual corrections, but be careful about asking third parties to make claims on your behalf, which may bring their content under your responsibility in some regimes. Ask compliance before running outreach in regulated sectors.

What should you log?

Keep a record of every published claim, its source, the reviewer and the review date. When an AI answer misstates something, you can show what you published and when. The log also tells you which pages need review when a rate, fee or guideline changes.

How do you respond to a harmful AI answer?

If an answer gives dangerous advice using your name, such as a wrong dosage or a misleading financial claim, treat it as a high-severity issue. Record it with a screenshot and date, check your own pages for the source, fix anything wrong, and use the engine's feedback route if one exists. Involve legal and compliance early.

What about generative tools inside your own product?

If your product uses AI to answer customer questions, the same rules apply to its answers, and sector regulators may have specific expectations. That's a separate topic from GEO, but your compliance team will likely ask about both together.

How do you train writers?

Give writers in regulated sectors a short guide: which claims need conditions, which words to avoid, which disclosures are required and how to cite sources. Include before-and-after examples of answer-first sentences that keep the required wording. Review the guide each year with compliance. Keep the examples current as rules change. Ask the reviewer to confirm the update. Record who signed off, when they did, and which version of the rule they checked it against.

Common questions

Does GEO change compliance rules? No. Content written for AI answers is subject to the same rules as other public content.

Can I write answer-first in regulated sectors? Yes, if the key condition or warning sits with the claim.

Who should approve GEO pages? Your existing compliance, legal or clinical reviewers.

Sources

All accessed 31 Oct 2026.

Questions

Can regulated firms do GEO?

Yes. GEO content is still marketing, so the same rules apply. For EU investment firms, MiFID II says marketing must be 'fair, clear and not misleading' and 'clearly identifiable as such'.

Who is responsible if an AI engine misquotes my regulated content?

This guide can't answer that legally. Our view: control what you can, which is your source pages. Make every claim on them compliant and unambiguous out of context.

Should regulated firms block AI crawlers?

Not by default. Blocking search crawlers removes your compliant content from AI answers, leaving engines to rely on other sources about you.