It’s understandable why there are so many articles on AI lately. Consumers increasingly ask ChatGPT, Claude, Gemini and other AI platforms questions they once typed into Google. So the new popular acronym – AEO (answer engine optimization) – for optimizing web content around AI engine results and recommendations is addressing a very real concern.
But after running a small experiment, we’re probably making AEO more complicated than it currently is.
I asked seven major AI platforms essentially the same question: “Who do you recommend to get a DSCR (debt service coverage ratio) loan from in Knoxville, Tennessee?”
I deliberately chose a product and market where I should theoretically have been competitive. I actively originate DSCR loans, and I operate a website specifically devoted to rental property financing.
I wasn’t in any of the seven engines’ recommendations. I wasn’t surprised I lost, but I was surprised by who beat me, and how familiar their tactics were. Some platforms surfaced large national investment property lenders. Others found brokers, specialty lenders or smaller lending websites.
But many businesses being recommended didn’t have any connection to the Knoxville area. They just had done something SEO-minded web designers have been doing for well over a decade.
They had published pages explicitly targeting searches like: DSCR loans in Knoxville, Tennessee. Then another for Nashville. Another for Chattanooga. Change the geography, adjust the page and repeat into the dozens, or hundreds.
These companies did not appear to be winning primarily because AI had conducted a sophisticated comparison of reviews and career experience. They had made themselves exceptionally easy to retrieve for the exact question being asked.
Retrieval comes before reputation
When I asked the models why they had not recommended me, several offered familiar advice: build backlinks, collect reviews, strengthen third-party profiles, publish authoritative content and maintain consistent business information.
All sensible. But those explanations didn’t fully account for their own first-pass results — and first pass is where the consumer stops.
That led me to what may be the most useful takeaway from the experiment: recommendation has two gates.
The first gate is: Can the machine find you? If it can, does it have enough evidence to recommend you?
One AI platform initially told me it could find little connecting my name with DSCR lending in Knoxville. When I supplied additional context — that I was a mortgage broker and operated a rental-finance website — it found the site, evaluated it and considered it relevant.
The problem was not necessarily that the AI had compared me with competitors and rejected me. It appears never to have evaluated me at all.
Individual loan officers have an identity problem
In only a couple of instances did an individual loan officer show up for a Knoxville DSCR recommendation. One Knoxville-area broker was an exception.
His profile page on his sponsoring company site made it easy to understand why. The content connected his name with investor lending, DSCR and fix-and-flip financing, Knoxville and East Tennessee, short-term rentals, his NMLS information and his contact details.
That matters because mortgage professionals face a problem that Realtors often do not.
My test on a natural Realtor-oriented question, “Who is the best Realtor in Knoxville?” immediately brought individual Realtors rather than brokerages. The question itself drives the engine toward individual practitioners.
Mortgage prompts are often phrased differently: “Who should I get a mortgage from?” or “Who is the best DSCR lender in Knoxville?” Those questions naturally point toward organizations. This broker’s unusually detailed company profile offers a clue as to how an individual originator can still break through.
His company’s level of detail was critical. Consider the standard mortgage biography on a company website: “John Doe is an experienced mortgage professional passionate about helping clients achieve the dream of homeownership.”
There is nothing wrong with that sentence. There is also very little in it that helps an AI system decide when John should be recommended.
For individual originators, that creates a problem. Our strongest digital authority lives in places we don’t control: employer websites, branch pages, company profiles, licensing records and third-party platforms.
That creates an opportunity for sponsoring organizations.
Many mortgage companies may be underusing some of their most valuable digital real estate when individual LO pages consist of little more than a headshot, phone number, NMLS number and interchangeable biography.
If an originator specializes in DSCR, say DSCR. Same for VA borrowers or first-time homebuyers. If she serves Knoxville and East Tennessee, say Knoxville and East Tennessee.
Do not make a machine infer expertise that can simply be stated.
Before optimizing for AI, fix the web you already have
This was an informal experiment involving one practitioner, one product and one metropolitan area, but the practical lesson seems simple. AEO is real. Consumers are already asking machines who they should call.
But traditional SEO fundamentals haven’t suddenly become obsolete because the consumer interface changed from ten blue links to a conversational answer. Far from it: at least in my testing, they’re remarkably important.
The newer challenge for individual practitioners is making sure authoritative pages clearly explain who they are, where they work and what they actually know how to do.
And mortgage company leadership should recognize that the humble loan officer profile page may now be far more valuable than it was even a few years ago.
AI does not just need to find the company. It increasingly needs enough information to understand the person inside it.
Morgan Hardy, AMP, is a mortgage broker based in Knoxville, Tennessee, specializing in real estate investor financing and DSCR loans. He provides guidance on investment-property lending at rental-mortgage.com.
This column does not necessarily reflect the opinion of HousingWire’s editorial department and its owners. To contact the editor responsible for this piece: [email protected].
