ChatGPT, Perplexity and Google's AI Overviews now recommend loan officers by name. Most mortgage sites were never built to be read by them — so they are simply left out of the answer. Find out where yours stands.
Free · Takes about 10 seconds · Full report emailed to you
This score reflects whether AI can read and understand your site — the foundation. Actually getting recommended also depends on your reviews, reputation, and how well your content answers what borrowers are asking.
We’re sending the complete write-up for to — every finding above in plain English, why each one matters, and the fixes ranked by what actually moves you into the answer.
When someone asks an assistant for a lender in your city, it returns a short answer with a few names in it. There is no page two. You are either one of those names or you are invisible — and what decides it is not what used to decide rankings.
Assistants often fetch pages without running your JavaScript. If your rates, programs and contact details only appear after a script loads, they can see an empty page — however good it looks in a browser.
A model has to place you: licensed lender, these loan products, these states, this NMLS. Sites that never state it plainly get passed over in local answers for competitors who do.
Models lean on corroboration — reviews, consistent listings, named authors, content recent enough to still be true. Thin, stale, anonymous pages give them nothing to stand on.
The first six are measured live the moment you hit Analyze — we fetch your site the way an assistant does, with JavaScript switched off. The last two need a person, and land with your emailed report.
Whether GPTBot, PerplexityBot, ClaudeBot and Google-Extended are allowed or quietly blocked in robots.txt.
What an assistant sees with JavaScript switched off — the copy that actually reaches the model.
Whether your name, NMLS, licensed states and loan products are stated in text a model can quote.
Organization, LocalBusiness, FAQ and review schema — present, valid, and matching the visible page.
Whether you answer the questions borrowers actually ask, in a form that can be lifted into an answer.
Name, address, phone and licence details agreeing across the places models corroborate against.
Dated, attributed content versus undated pages with no named author behind them.
The same Lighthouse pass Google runs — still the floor everything else is built on.
Every mortgage site we launch ships server-rendered, schema-complete and fast by default — the things this audit looks for, already done. Pick a design and we build it in three days.
Manage how we use cookies. Necessary cookies keep the site working and can't be turned off.
Saved on this device. Press Esc to close.