Marketing

How to Get Your Firm Recommended by ChatGPT and Perplexity

Getting named inside AI answers is now a real client-acquisition channel for advisors. Here is the research-backed playbook: what ChatGPT and Perplexity actually read, and the five moves that get a firm recommended.

Bhavya Barot

Bhavya Barot

— 8 min read
Illustration of an advisor profile, public source cards and an AI answer with references.

Start by making your firm's public facts complete and consistent, publishing answers to the questions you can help with, and checking how ChatGPT and Perplexity describe the firm over time. Record mentions, citations and errors separately. You can improve the information available about your firm, but you cannot ensure an AI service will recommend it.

We maintain a step-by-step version of this playbook here: How can an RIA get clients from ChatGPT and AI search?

Key takeaways

  • Ahrefs found correlations in selected AI-visibility samples, not a recommendation formula or client-acquisition guarantee.
  • ChatGPT assembles its answer from your site, directory profiles, press, podcasts, and reviews - not from any ad you can buy.
  • First step: complete, consistent public facts plus content that answers real client questions in a form an AI can lift and cite.

On this page

We spend our days inside this question at Valora, so we read the research seriously. Three findings do most of the work in this piece, and they are worth your time even if you never read anything else on the topic.

Finding one: AI assistants lean hard on what the web already says about you

Ahrefs' May 2025 analysis of 75,000 brands found a Spearman correlation of 0.664 between web mentions and Google AI Overview visibility in its selected sample. Its December 2025 follow-up found YouTube mentions had the strongest measured correlation, about 0.737, across ChatGPT, AI Mode and AI Overviews. These are observational associations, not causal levers, recommendations or client-acquisition results. Sources: May 2025 Ahrefs study; December 2025 follow-up.

Public facts and independent references are material a reader or search system may encounter. Keep them accurate and consistent without treating them as a disclosed recipe for ChatGPT recommendations.

Finding two: AI answers cite sources, and you can become one of them

The GEO research presented at KDD 2024 tested content changes on its GEO-bench query benchmark. It reported source-visibility gains of up to 40% under its metrics, with results varying by domain. This does not establish an advisory-firm client-acquisition effect or reveal the algorithm of a commercial AI tool. Sources: GEO paper.

Notice what those tactics have in common. They make your content look like evidence. A page that says "we provide holistic wealth management" gives an AI nothing to work with. A page that says "43 percent of our new clients arrive holding concentrated employer stock, and here is how we handle the first 90 days" gives it a quotable, specific, checkable claim. The AI is looking for the second kind of sentence.

Finding three: the volume of AI answers is large enough to matter

Ahrefs' May 2025 dataset contained about 55.8 million AI Overviews observed in desktop results across 590 million keywords in its index. Those are keyword-result observations, not 590 million individual consumer searches, and do not establish lost clicks or advisor hiring behavior. Sources: Ahrefs desktop-keyword analysis.

Put the research together and the playbook is not mysterious. It has five parts.

1. Decide what you want to be recommended for. "Financial advisor" is too broad to win. The AI assembles answers to specific questions: advisor for physicians, advisor for equity compensation, advisor for retirement income in a specific city. Pick the two or three questions your best clients would actually type, and build your visibility around those.

2. Say the same thing everywhere. Your website, your Google Business profile, your LinkedIn, your directory listings, and your ADV should describe the same firm, serving the same people, in the same words. Inconsistency reads as uncertainty, and uncertain entities do not get named.

3. Publish pages that answer real questions with real substance. Not keyword pages. Pages that genuinely answer "how does a 401(k) rollover work when you leave a tech company" with specifics, numbers, and your point of view. Per the GEO research, cite credible sources, quote real voices, and use statistics. Be the source the AI quotes.

4. Review relevant outside mentions. Keep controlled directory facts current and contribute only supportable expertise to suitable publications or events. Record mentions separately from citations, inquiries and clients.

5. Make your site machine-readable. Clear page structure, descriptive headings, FAQ sections, schema markup where it fits, and no important text trapped inside images. The AI can only recommend what it can read.

The part nobody wants to hear

Useful public information takes ongoing maintenance. Neither the timing nor the outcome of AI visibility is established for your firm; use repeated dated checks rather than a promise of owning answers.

Request a review of your firm's public website. We email a short, dated report with observations, limits and next steps; it is not an instant score or an AI recommendation guarantee. Sources: Request a public-website review.

ChatGPT and Perplexity do not read the same web

LumenGEO, a vendor, reported a July 2, 2026 snapshot using 130 matched queries and 3,292 citations across Perplexity and ChatGPT's web-search-enabled Responses API. The observed source sets differed. This is not a consumer-UI study, proof of how either product's algorithm works, or a forecast for an advisory firm. Sources: LumenGEO vendor snapshot.

The practical consequence: you cannot optimize for one engine and assume the others follow. The safe strategy is the boring one - be mentioned widely, be described consistently, and publish content worth citing. That portfolio works on every engine, including the ones that do not exist yet.

Illustration of two fictional advisory offices with contrasting public information displays and a question connecting to source cards.
A hypothetical comparison of how easily public firm facts can be checked; no firm or AI system is shown as a winner.

A worked example: two firms, one question

Imagine two RIAs in the same city, similar size, similar credentials. A prospect asks Perplexity: "financial advisor for doctors in Denver."

Firm A's website says it offers "comprehensive wealth management for individuals and families." Its directory profiles were last touched in 2022. It publishes nothing. Nothing else on the web mentions it by name.

Firm B's website has a page titled "Financial planning for physicians in Denver" that walks through residency-to-attending transitions, disability insurance, and student loan strategy with actual numbers. The founder has been quoted twice in trade press about physician finances. Three podcast appearances carry show notes with the firm's name and specialty. Its profiles are complete and identical everywhere.

In this fictional comparison, Firm B gives readers more specific, consistent information to inspect. That could help a search system describe it accurately, but neither firm is assured a mention, citation or recommendation.

What does not work

Avoid unsupported claims, fabricated reviews and near-duplicate pages that do not serve distinct questions. Use truthful, reviewed material rather than implying knowledge of a model's hidden recommendation process.

Illustration of question cards, a calendar, a magnifying glass, and five stage icons for a page, document, message, calendar and folder.
Track each stage separately; a mention is not an inquiry, meeting or client.

How to measure progress

Use a fixed set of questions for a dated visibility test. Record the prompt, model, response and citations. A citation can suggest a source to inspect, but it does not tell you exactly what to publish next or establish a ranking recipe.

Done researching? The free readiness check shows whether AI search can find and recommend your firm, and you can reach the team straight from it: Can ChatGPT find your firm?

The honest summary

Getting recommended by AI is not a trick. It is the same reputation work good firms have always done - publish useful things, be specific about who you serve, show up in places that quote you - with one new requirement: the evidence has to exist in writing, in public, in a form a machine can read. Advisors who built their practices on handshake referrals now need that same reputation to exist on the web. Start before your competitors figure that out.

A 90-day starter plan

Days 1-14: fix the foundation. Rewrite your homepage and bio to name exactly who you serve and what you do for them, in plain language. Align every directory profile, your LinkedIn headline, and your Google Business profile to the same description. This is the consistency layer everything else sits on.

Days 15-45: publish three answer pages. Pick the three questions your best clients asked before they hired you. Write a full page on each, with specifics, numbers, and your firm's actual point of view. Cite credible sources where you make claims - the GEO research says citations and statistics are what make content quotable to a machine.

Days 46-90: seek relevant outside mentions. Choose suitable publications, podcasts or events and offer supportable expertise. Five mentions is a planning target, not an achievable result promised to every firm.

Repeat dated checks and compare the same questions. Ninety days is a working review period, not a promise of visibility, dominance or being the only firm named.

Questions advisors ask us about this

AI tools are one possible research resource alongside referrals, search and other sources. A mention does not ensure a referral becomes an inquiry; absence does not establish a lost client or unspoken doubt.

There is no established timetable for consistent appearances. Record prompt, date, model and observed results; assess changes without treating the correlation studies as evidence that mentions cause visibility.

Is this different for small firms? It favors them, for now. A solo RIA with a tight niche and five good answer pages can own a specific question outright, because the big firms publish generic content at scale. Specific beats big inside an AI answer.

Sources and scope: May and December 2025 Ahrefs correlation samples; the May 2025 desktop-keyword dataset; the KDD 2024 GEO benchmark; and the LumenGEO July 2, 2026 vendor snapshot. Each answers a different research question, not whether your firm will win clients. Sources: Ahrefs May; Ahrefs December; Desktop results; GEO; LumenGEO.

Where this advice has limits

Results vary by prompt, model and date. Correlation does not establish that adding mentions causes a citation, and visibility is separate from a suitable inquiry or signed client.

Where Valora fits

Valora is testing ways to help advisers understand public information and measure the path from visibility and outreach to conversations. It cannot ensure recommendations. A requested review produces a short, dated report by email. Sources: Request a public-website review.

The bottom line

Keep firm facts accurate, publish useful supportable answers and check what each AI service returns. Record citations, errors and real inquiries separately; no recommendation is assured.

Frequently asked questions

1. How does ChatGPT decide which financial advisors to recommend?

A ChatGPT answer can use retrieved public sources, but this article does not disclose its ranking or recommendation algorithm. Ahrefs' May 2025 analysis of 75,000 brands found a Spearman correlation of 0.664 between web mentions and Google AI Overview visibility in its selected sample. Its December 2025 follow-up found YouTube mentions had the strongest measured correlation, about 0.737, across ChatGPT, AI Mode and AI Overviews. These are observational associations, not causal levers, recommendations or client-acquisition results. Sources: Ahrefs May study; December follow-up.

No. There is no ad slot inside a ChatGPT answer. Visibility comes from being mentioned and cited across the public web in ways AI engines can verify.

3. What is the first step to showing up in AI answers?

Make your firm's public facts complete and consistent - who you serve, what you do, where, credentials, fees - then publish content that answers the exact questions your ideal clients ask, in a form an AI can lift and cite.

Want to see whether AI search can find and recommend your firm? Run the free readiness check - and if you want to talk it through, you can reach us straight from the results: Can ChatGPT find your firm?

Read more