AI Citation Monitor
AI visibility for financial advisors, measured every month.
Google shows ten results. An AI answer names three to six advisors and stops. AdvisorSEO Max is the advisor-specific instrument for knowing whether you are one of them: a fixed question set built for your metro, hundreds of scored answers a month across Claude, ChatGPT, and Perplexity, and a report your compliance reviewer can read as is.
Start your 14-day free trial Book a 20 minute demo14-day free trial. Cancel anytime. Included in the one plan, $199 per month.
AI Citation Rate
54%of 368 responses
199 of 368 responses cited my site. 230 named my firm.
- Perplexity 68% of 128
- ChatGPT 55% of 112
- Claude 39% of 128
- Broad discovery 51% of 92
- Niche discovery 91% of 92
- Brand recognition 100% of 92
Melby Wealth Management, September 2026, first measured month. One firm, mine, over a defined window. An observation, not a projection for yours.
There is no page two.
When a prospect asks an AI assistant for a fee-only advisor in your city, the answer names three to six firms and stops. There is no position eleven to climb from. A firm that is absent from the answer does not exist for that prospect. The answer also changes by assistant, by run, and by where the asker is sitting, which is why typing the question into ChatGPT yourself once tells you almost nothing.
AI answers were the second-largest source of new prospects at Melby Wealth in 2026 through August, behind Google. That is the channel this page is about.
AI answers are built from a different set of sources than Google ranks: directory profiles, review sites, local media, and competitor pages. A firm can hold a page one position for "financial advisor" in its city and be missing from the chat answer, because the chat answer was assembled from pages the firm was never on. Measuring the answer, month after month, is the only way to know which case you are in.
AI search optimization for advisors, measured instead of guessed.
Answer engine optimization, AEO, is the work that makes a firm citable: schema that says who you are and where you practice, headers and answer-first pages a model can lift from, consistent firm details across the web, and profiles on the directory pages the models already trust. None of that is exotic. What has been missing is a way to know whether it worked.
That is what the monitor is for. Do the AI search optimization work, then watch the citation rate, per assistant, with its sample size beside it. When the broad discovery rate moves after you complete two directory listings, you know which lever moved it. When it does not move, you know that too, before you spend another month on the wrong thing.
One spot check is an anecdote. Three hundred is a rate.
20
questions per firm, built for your metro: 16 discovery questions across four types, plus 4 brand checks scored on their own line.
3
assistants measured through their own APIs with web search on and your city passed as the asker's location: Claude, ChatGPT, Perplexity.
368
discovery answers read and scored in my first month, plus 92 brand-check answers. Monthly on the 15th, plus one on-demand run.
Per assistant, with the model version measured.
The gap between assistants is the finding most advisors never see. Perplexity cited my site nearly twice as often as Claude did in the same month, on the same questions.
All providers
54%
199 of 368 responses cited my site. Named in 62%.
368 scored responses
Claude
39%
50 of 128 responses cited my site. Named in 40%.
claude-sonnet-5
ChatGPT
55%
62 of 112 responses cited my site. Named in 69%.
gpt-5.6-luna
Perplexity
68%
87 of 128 responses cited my site. Named in 80%.
perplexity/sonar
Melby Wealth Management, September 2026. Rates carry their sample size because answers vary run to run. When a provider swaps models, the report says so and pauses that provider's comparison.
A strong niche rate can hide a weak broad rate.
Every question is scored by type, and the broad discovery row sits first because it is the hardest to win and the one worth judging yourself by. Here is my own first month, cited rate by question type.
Melby Wealth Management, September 2026. Brand checks measure recognition, not discovery, so they are tracked separately and never lifted into the headline rate.
The pages AI trusts instead of yours, and whether you are on them.
Every page the models cited across the month, with a status. "Not listed" means the monitor fetched the page and did not find your firm on it. Sorted by how often the models trusted them, those rows are the to-do list.
| Page cited | Responses citing it | Status |
|---|---|---|
| melbywealth.comhomepage | 286 of 368 | Your site |
| smartasset.comnashville-tn-top-financial-advisors | 154 of 368 | Not listed: pursue |
| napfa.orgfirm profile | 138 of 368 | Listed |
| Competitor firm sitenamed in your report, withheld here | 134 of 368 | Competitor site |
| melbywealth.com/financial-planning | 129 of 368 | Your site |
| feeonlynetwork.comnashville-tn-financial-planner | 115 of 368 | Listed |
| melbywealth.com/services | 82 of 368 | Your site |
| expertise.comfinancial-advisors/tennessee/nashville | 78 of 368 | Not listed: pursue |
Rows from my own September report. Competitor firm names appear in your report and are withheld on this page.
The questions your competitors win every time.
Each question gets its own line: how many answers cited your site and how many named your firm, per assistant, with a real sample answer and the sources it used.
Two of mine came back almost empty. Asked who provides investment management for individuals and families in Nashville, competitor pages were cited in 19 of 23 responses and mine in 2. Asked whether someone with about $1 million should use a large wealth management firm or an independent RIA, competitors were cited in 22 of 23 and mine in 4.
Both fixes are the same shape: one page that answers that exact question. The October cycle grades the work.
Where the same month went right
The question about which Nashville firms focus on serving a millennial came back cited in 23 of 23 responses. The musician question, 20 of 23. The plain "who are the best fee-only financial advisors in Nashville" question, 19 of 23.
And the competitive read
In broad discovery, one competing firm was cited in 55% of 92 responses against my 51%. That gap is the kind of thing you cannot see without measuring, and it is the reason the report puts the broad row first.
Every number ships with its sample size and its model version.
Named and cited are reported separately
Being mentioned in an answer is a different thing from being used as its source. Both rates appear, and neither is folded into the other.
Brand checks never inflate the headline
Questions that ask about your firm by name are tracked on their own line. A model knowing who you are is not the same as a model recommending you.
Broad discovery gets its own line
The plain "find me an advisor in my metro" questions are the hardest to win and the most valuable, so their rate sits directly under the headline where a strong niche rate cannot hide it.
Model changes are flagged, not hidden
Each assistant's rate carries the exact model version measured. When a provider swaps models, the report says so and pauses that provider's comparison.
An arrow always means something
Month-over-month change appears only when the prior month's sample was large enough and no model changed underneath.
No rank, no "best," no score out of 100
Rates with denominators, the models' own answers quoted as evidence, and a SWOT where every line carries its numbers. Built to be forwarded to a compliance reviewer as is. Whether they like what they read is their call.
Your first report, within the hour.
The baseline cycle runs the moment the monitor is turned on and publishes within the hour. Your first login shows your citation rate and the pages you are missing from.
On the 15th, whether or not you log in.
A six-page branded PDF arrives by email every month with the same content as the dashboard. Commander answers "am I showing up in Claude?" from the same data, and your Citation Monitor scores appear in the Monday email alongside everything else that changed.
Questions advisors ask
AI answers change every time. How can a rate mean anything?
That variation is the reason to measure. One spot check tells you nothing, so each question is asked repeatedly per model per month and reported as a rate with its sample size next to it.
Is this the real ChatGPT people use?
It is the same models, measured through each provider's API with web search enabled and the asker's location set to your metro. That is the only way to measure consistently at scale, and every report says so in its methodology note.
My rate is low. Now what?
The report is built to answer that: which question types are weak, which directories the models cite that you are not on, which specific questions competitors win, and what the models say when asked about you by name. Each one is a task an advisor or their marketing person can do this month.
Can you guarantee I will show up?
No, and no honest tool can. The monitor measures and points at the specific listings and pages that close the gaps. The rate is the scoreboard.
Is this compliant?
The product makes no ranking or superiority claims, quotes the models' own answers as evidence, and reports rates with denominators. It is compliance-aware by design. Run your marketing through your own compliance review as usual; nothing here certifies anything.
What about Google AI Overviews?
There is no direct check for Google AI Overviews. AdvisorSEO Max scores the structural foundation Google AI Overviews reads (schema, headers, consistent firm data) and the same on-page work carries there.
I would rather have this done for me.
AdvisorSEO Max Managed runs the platform on your site and does the implementation, one firm per market, three firms at a time. Details are on the Managed Service page.
Included in the one plan. $199 per month.
Everything in AdvisorSEO Max, no feature gates, no add-ons. The AI Citation Monitor, the on-demand AI Visibility Check, Verified Wins, the compliance review scanner, Commander, and the Form ADV tools.
Start your 14-day free trial Book a 20 minute demo14-day free trial. Cancel anytime. The demo shows the monitor on my own firm's data.