AI assistants don’t keep a list of good advisors. When a client asks ChatGPT, Claude or Perplexity “Who’s a good fee-only advisor in Denver for someone with stock options?”, the assistant usually runs a web search, reads a handful of pages, and names the firms those pages describe clearly. Whether your firm comes up depends on three things: whether the assistant can reach pages about you, whether those pages answer the client’s question, and whether the facts agree with each other.
What happens when a client asks
The client’s question is long and specific. People talk to AI the way they’d talk to a friend: their age, the move, the stock options, the inheritance, what worries them. The assistant turns that into a few searches, reads the results, and writes an answer that names a few firms or advisors, often with a link.
Three kinds of pages tend to feed that answer:
- The firm’s own website, when it states plainly who the firm serves, where, and what it charges.
- Directories and profiles that list advisors by place and specialty, especially ones that show fees and credentials.
- Articles and press that mention an advisor by name on the topic the client asked about.
The assistant doesn’t need to know your firm in advance. It needs to find a page that matches the question and trust it enough to repeat it.
The bots behind the answers
Each AI company runs different crawlers, and they do different jobs. Some build the search index an assistant draws on later. Some fetch a page in the moment, because a person asked a question. Some collect pages to train future models.
Together they add up. In the 28 days to October 30, 2026, AI assistants and AI search requested Wellspent’s advisor pages and data 38,670 times, counting only verified requests from the AI companies. Most of that came in the last few days, as AI search engines indexed the full public graph.
The in-the-moment fetches matter most for advisors. Each one is an assistant opening a page because a real person asked a question that page might answer. The search indexers decide which pages are available to be found later.
One warning if you look at your own analytics: much of the traffic that claims to be an AI bot isn’t. Anyone can put an AI company’s name in a request. Count only traffic your host verifies as coming from the AI company.
When we asked ChatGPT and Perplexity for fee-only advisors in 100 cities, every answer named specific firms, about 8 on average, but the two assistants agreed on fewer than 1 in 5 of them for the same city. Most of the difference comes from what each one reads. The full results are in our study of 600 AI answers.
The questions clients actually ask
Search engines trained people to type three words. AI assistants get whole situations. The questions that lead to an advisor recommendation tend to look like these:
- “I’m 58, my wife and I have about $1.8 million saved, and I want to retire at 62. Who should I talk to in Austin?”
- “My company just went public and I have RSUs vesting. I need someone who understands taxes on stock comp, not a salesperson.”
- “I’m going through a divorce in Phoenix and need help deciding whether to keep the house.”
- “My dad died and left me an IRA. Do I need a financial advisor, and how do I find one who doesn’t earn commissions?”
Each one carries a place, a situation, sometimes an amount, and often a preference about how the advisor is paid. The assistant searches for those pieces. A firm that writes about pre-retirees in Austin, equity compensation, divorce, or inherited IRAs on its own pages has text that matches. A firm whose pages describe “holistic wealth management” doesn’t.
The questions also tell you who these prospects are. People who describe their whole situation to an AI assistant before calling anyone tend to arrive informed, with specific needs and a short list.
Why some firms never come up
When a firm doesn’t come up, it’s usually for one of a few reasons.
The facts aren’t on the page
A homepage that says “We help you pursue your financial goals” gives an assistant nothing to match. If a client asks about RSUs, the firm that writes “we plan equity compensation for tech employees” on its site can be matched. The firm that only says “comprehensive wealth management” can’t. In our website study, about 4 in 10 fee-only firm websites named no specialty beyond retirement, tax or estate planning.
The price is missing
Clients ask what advice will cost. An assistant can only answer with a number it finds. About half of fee-only firm websites list any price we could verify, so the firms that do publish a fee get quoted.
The facts disagree
If your website says one city, your directory listing says another, and an old profile lists a former firm, an assistant has to guess. Conflicting facts make it less likely to name you at all, and more likely to state something wrong when it does.
The page can’t be read
Some sites block AI crawlers in their robots.txt file, often by accident through a security plugin or hosting setting. Others put key facts inside images, PDFs or scripts the crawler doesn’t run. If the crawler can’t read it, it can’t repeat it.
Nothing outside your site backs you up
Assistants lean on independent sources: directories that verify credentials, press mentions, profiles that agree with your site. A firm that exists only on its own website is easy to overlook, however good that website is.
When AI gets your firm wrong
Being named isn’t the only risk. Assistants sometimes state things about a firm that aren’t true: an old address, a former partner, a minimum the firm dropped years ago, or “fee-based” for a firm that is fee-only. These errors almost always trace back to a page somewhere that says it, such as an outdated directory profile, an old press release, or a cached version of a previous website.
You can’t edit the answer, but you can fix the sources. Search for your firm name and your own name, open every profile and listing that comes up, and correct or remove what’s out of date. Make sure your current website states the facts that are most often wrong, in plain text, near the top of the page.
Why big firms show up so often
Large firms appear in AI answers more than their size would suggest because they have more of everything an assistant reads: more pages, more press, more directory listings, more mentions. That’s a volume advantage, not a quality one. For a specific question (“fee-only planner for a divorce in Phoenix”), a small firm with clear, matching facts on the right pages can be named ahead of a national brand. The more specific the question, the more the facts matter and the less the brand does.
What one firm saw
On a March 2026 episode of Kitces’ Financial Advisor Success podcast, Helen Stephens described building her firm to $660 million in assets largely through search, and then seeing rising traffic and near-automatic inquiries from ChatGPT. Two points from her account match what we see:
- What converted shifted over time from geography to fee-only status to detailed answers on tax questions. Specific, answer-shaped content kept working as search changed.
- AI assistants tend to send people to a human advisor when a question gets complex. The prospects AI sends you are often the ones with the most complicated situations.
How we measure it
Wellspent members get a monthly AI visibility report. We write the questions a member’s ideal clients would ask, based only on what the member has published: their specialties, location, fees and the clients they serve. We put those questions to an AI assistant with web search on, the way a client would. Then we pull out every advisor and firm the answer names, and the sources it cited. The report shows whether the member was named, who was named instead, and which pages the answer leaned on.
We’re running the same method across 100 US and Canadian cities for a public study.
What to do about it
The short version: put clear, consistent, sourced facts about your firm on the pages AI reads, and let the crawlers in. The step-by-step version, with what to publish and where, is in How to get your firm recommended by AI assistants.