Loading Marketing Pulse…
Marketing Pulse

Most top producers are invisible when a buyer asks AI

Rates are flat and there is nothing new to react to, so this week's lever is discovery: new Local Falcon research says 91.5% of top-producing agents never appear in AI search answers, and the same gap is sitting under every loan officer's name.

Sunday, August 23, 2026 30Y 6.77%15Y 6.61%5/1 ARM 6.36%

Rate environment is quiet this week and there are no new lender, GSE, or regulatory moves to react to, so the marketing story is a discovery one. Local Falcon research reported by Inman found that 91.5% of top-producing agents never show up in AI search results at all, and among the ones who do surface, the citation frequently traces back to a "best agent" ranking page they published themselves. Roomvu shipped a free tool called Found the day after, which checks how a given agent or broker ranks across ChatGPT, Gemini and Claude and flags what to fix. The research is about agents, not loan officers, and it is worth being honest about that — but the mechanic is identical. When a buyer types "best mortgage lender in Boise" into a chatbot instead of a search bar, the answer is assembled from pages that mention you by name in a context the model can parse. If nothing does, you are not ranked poorly. You are absent.

Nothing in the rate picture is going to generate that call for you this week. Bankrate's 30-year has printed 6.72% for three straight sessions, it has traded between 6.67% and 6.82% for the last 30 days, and it sits eight basis points above its 90-day average of 6.64%. Principal and interest on a $400,000 loan runs about $2,586 a month, essentially where it was a month ago. There is no "rates just dropped" hook available and pretending otherwise costs you credibility the moment the borrower checks. The one segment with real math behind it remains borrowers still carrying 2023 paper above 7.5% — on a $400K balance that is roughly $190 a month, which is a conversation worth having even in a flat market. Everyone else on your list needs a reason to find you that has nothing to do with today's number.

The tactical move is to make yourself quotable. AI answers get assembled from text that states a fact plainly and attributes it to a named person in a named place, so the pages that win are the boring ones: a bio page that answers "what does an FHA loan require in Ohio" in full sentences under a heading shaped like that question; a local association or chamber directory listing with your specialty spelled out; a guest paragraph on an agent partner's blog that names you and your market. Reviews matter here too, but only the specific ones — "helped us with a VA loan in Tacoma with a 30-day close" is retrievable, "great to work with" is not. Ask your last ten closed borrowers for a review that names the loan type, the city, and the timeline, and you have built ten quotable sentences.

Do this today

open ChatGPT, Gemini, and Claude, ask each one "who is the best mortgage loan officer in [your city]," screenshot what comes back, and write down every named source it cites — that list is your outreach plan for the week.

Borrower segments to act on today

2023 paper still sitting above 7.5%

These borrowers closed at or near the cycle peak. Against today's 6.72% conventional, a 7.5%+ note is roughly $190/mo on a $400K balance, which clears break-even inside two years on standard origination costs even with rates flat.

closed loans · ≥24mo since close · rate ≥7.50%
Recent purchase closings — your review supply

Borrowers who funded in the last six months still remember the details and convert to reviews at the highest rate. Each specific review is a retrievable sentence an AI answer can cite; this is the list you mine to fix the visibility gap.

closed loans · ≤6mo since close · purchases

Today’s content angles

Short-form video

Ask the chatbot your own name

Face to camera: "I asked ChatGPT who the best mortgage lender in [city] is. I was not on the list. Neither was anyone I know who actually closes loans here. Here is what it named instead, and why that matters if you are shopping for a loan this fall. If you want a real person who will run your actual numbers, message me and I will send your payment on paper."

Tactics worth stealing

Write the page an AI can quote back

Retrieval favors plain declarative sentences under question-shaped headings, attributed to a named person in a named place. Rewrite one bio or FAQ page so each heading is the literal question a borrower would type and the first sentence under it answers that question completely, without a preamble. Vague brand copy is unretrievable no matter how well it reads.

Local Falcon AI search visibility research, reported by Inman (Aug 2026)