Salt Marsh Growth AI
Start with a Discovery Conversation

Real Estate

Where AI fits your brokerage — and where it doesn't.

We inspect how inquiries, follow-up and client communication work in your office today, then recommend what to keep, what to configure in the tools you already pay for, and what to change. Every recommendation is written and reviewed by a named person before you see it, and any estimate expressed in closings shows its assumptions.

A real estate agent showing a home to a couple.

What we look at

Whether AI names you, and what it draws on.

Ask ChatGPT, Claude, Perplexity, Google or Gemini “who are the best real estate agents in [your town]” and each writes an answer that may name firms. We record what those answers say on a stated date — who is named, what is said, and which pages are cited — and what your own reviews, profiles and published pages give the services to work from. It goes in the report as a dated finding.

What happens to the leads you pay for.

How an inquiry from Follow Up Boss, kvCORE, Real Geeks or your website reaches an agent, how long that takes at 7 p.m. on a Saturday, and what your records show happened next. Where a record shows why an opportunity was lost, the report says so; where it can only support an estimate, it says that, with the assumptions.

How past clients hear from you.

How past clients and referral sources hear from the brokerage now — who sends what, how often, and whether anyone can tell what came of it. If a recommendation would put a routine update on a schedule, it names who reviews it before it goes out.

What your agents know about the local market, and when.

Where your agents get their read on the local market — insurance repricing, submarket velocity, the places a county median hides several different markets — and how current that read is when a client asks. The report says what is already in hand and what would need a source.

Where AI is already in the office.

Which tools agents use for listing descriptions, client emails and market summaries; how that output is checked before it is used; and what the brokerage's written AI policy covers, if there is one. AI-written listings still need review for accuracy and fair housing, the same as anything an agent writes.

Where to start depends on what the inspection finds. The report puts its recommendations in order, says what each one rests on, and leaves the decision with you.

See how AI picks who to name →

Questions to consider before an AI Audit

Is our CRM already enough?
It may be. We look at what it supports and how follow-up is set up in it before we recommend anything beyond it.
Are we losing deals to whoever answers first?
We look at response times and what happened next in your records. Where a record shows why an opportunity was lost, the report says so; where it can only support an estimate of what a delay might have cost, it says that, with the assumptions.
Will this change how my agents work?
The inspection changes nothing; we look at how the work is done today. If a recommendation would change something, it names what, who reviews it, and who keeps it running.
My agents already use AI for listings and emails. What should I be watching?
Two things. How listing and email drafts are checked before they go out — for accuracy, for anything confidential, and for fair housing — and whether you have a written AI policy that names the approved tools, the permitted uses, and who reviews the output. NAR offers a broker policy template as a starting point (member login required).
I already pay for SEO. Is it doing anything?
We read your search and inquiry records alongside the dated AI answers. The report separates what those records show about where inquiries came from and what would need further measuring.

Start with a Discovery Conversation

30 minutes. We learn your business; you learn how we’d work with you. No pitch deck. No obligation.

Start with a Discovery Conversation