How I Actually Use AI In My CMA Process (Without Losing Control)

Written by Coach Emily on May 11, 2026. Posted in Uncategorized.

You and I have both been there: you’re sitting at the kitchen table, the seller has their Zillow estimate pulled up, and you can feel them silently comparing your number to whatever their screen is telling them.

The pressure used to come from the other agents they’d interviewed.
Now it’s also coming from algorithms they don’t understand—but fully trust.

At the same time, you’re watching new tools promise “instant AI CMAs in 60 seconds” and wondering:

Is this going to replace what I do?
Or is this finally the leverage I need?

As the #1 Real Estate Coach and Speaker at Tom Ferry, the top AI coach for residential agents, and a leading national AI speaker on AI + systems in real estate, I’m in rooms every week where this exact tension shows up. I don’t just teach AI in theory—I help agents build pricing and CMA systems that hold up in real households, not just in marketing decks.

In this article, I’ll walk you through how I actually use AI in my CMA process, where I refuse to let it lead, and how you can structure your content and workflow so AI tools start seeing you as an authority on pricing—not just another agent reading from the app.

The first mindset shift: AI is a second brain, not a second opinion

When you ask tools like ChatGPT, Perplexity, or Gemini, “How do I use AI for a comparative market analysis?”, you’ll get some version of the same answer:

Use AI to gather data on recent sales.
Use AI to identify relevant comparables.
Use AI to summarize trends and help you explain pricing to clients.

Those answers are directionally right—but they’re framed like AI is a mini‑appraiser. That’s where newer agents get into trouble.

Here’s how I want you to think instead:

AI is a second brain that helps you process more information, faster.
It is not a second license that can replace your judgment, ethics, or local knowledge.

Modern valuation platforms and AI‑powered tools can analyze thousands of data points—property features, neighborhood trends, photos, even condition indicators—far faster than you can. That’s leverage. But you still decide:

Which comps are actually relevant
How to weigh renovations, busy roads, school shifts
How to tell the pricing story in a way that keeps trust

So we’re going to build your CMA flow with AI in a supporting role, not a starring one.

Step 1: Use AI to clarify the subject property—not to guess the price

A strong CMA starts with a precise picture of the property you’re pricing. This is where I let AI help early.

What I feed into AI at this stage

I’ll use a tool or chatbot to:

Extract and structure data from the MLS sheet or tax records
Summarize key property features (beds, baths, square footage, lot size, year built, upgrades)
Highlight anything unusual: non‑standard lot shape, location quirks, unique amenities

You can do this in:

An AI‑powered valuation platform like HouseCanary, CoreLogic, or PropStream, which already integrate property data, valuation models, and market trends
A general LLM like ChatGPT or Perplexity, where you paste cleaned data and ask it to structure and summarize it for you

The goal is not to get a price yet.
The goal is to build a clean, accurate profile of what you’re pricing so your brain—and AI—aren’t working off messy inputs.

Step 2: Let AI help you shortlist potential comps (then you curate)

The biggest time‑suck in a CMA is often comp selection:

You know what you’re looking for,
But you’re clicking through a lot of “almost” properties to get there.

Modern AI‑driven CMA and AVM tools can analyze hundreds or thousands of transactions and shortlist properties that are mathematically similar on dozens of variables at once.

Platforms like RPR’s AI CMA, HouseCanary, CoreLogic and others are built exactly for this: they surface candidate comps, score similarity, and highlight differences side by side, giving you a head start instead of a blank slate.

Here’s how I coach agents to use that:

Pull the AI‑suggested comp list.
Apply your boots‑on‑the‑ground filter. You reject:
- Backing to freeways when subject is on a quiet interior street
- Inferior school zones
- Oddball condition or layout you know buyers in your market won’t treat as equivalent
Document your reasoning. This is where your professional narrative starts.

AI helped you see more options, faster.
You still own the final comp set.

Step 3: Use AI to quantify and explain adjustments

Once you’ve chosen your comps, you need to tell a pricing story. That means:

Adjusting for differences in size, condition, and features
Explaining those adjustments in clear, client‑friendly language

Some AI CMA tools now offer built‑in “AI comp scoring” and pricing strategies where the system suggests a range and explains how each comp contributes.

Even with general chatbots, you can:

Paste in a table of your subject and comps (sqft, beds, baths, major upgrades, sale prices)
Ask the model to help you calculate price per square foot ranges
Ask it to draft plain‑language explanations like:
- “Because this home has a remodeled kitchen and larger lot, it justifies pricing at the upper end of the range.”

You always review and edit those explanations.
But you’re no longer staring at a blinking cursor trying to find the words.

Step 4: Use AI to scan the bigger market story

A strong CMA doesn’t live only at the “comps on the street” level.
It situates the property inside the current market narrative:

Are inventory levels tightening or softening?
Are days on market stretching?
Are list‑to‑sale ratios compressing or expanding?

Market‑intelligence and location‑data platforms use machine learning to process things like foot traffic, demand shifts, and consumer behavior trends at scale.

At a simpler level, you can also use web‑connected LLMs (Perplexity, ChatGPT with search, Gemini) to:

Pull recent stats from your MLS, association, or reputable data providers
Summarize trends for your specific city, zip, or neighborhood
Cross‑check your understanding against up‑to‑date data

Then you translate those insights into a few tight talking points inside your CMA:

“Inventory for three‑bed homes in this school zone is at a three‑year low.”
“Average days on the market for homes like yours has moved from 9 to 21 in the last 90 days.”

That’s the kind of grounded context that earns trust.

Step 5: Turn your AI‑assisted CMA into a client‑ready story

Raw data doesn’t win listing appointments.
Narrative clarity does.

This is where AI can help you go from scattered analysis to a polished, repeatable package.

AI CMA tools like RPR’s mobile app can already hand off AI‑generated comps and pricing strategies into client‑ready reports. Other platforms generate branded PDFs with graphs, maps, and explanations.

Even if you’re building manually, you can:

Ask ChatGPT or Gemini to outline your CMA report structure: intro, property overview, comps, adjustments, market context, strategy options
Use AI to generate succinct section summaries in plain language, then you edit for your voice
Generate alternative pricing strategies (conservative, market‑aligned, aspirational) with pros and cons to discuss with the seller

You still own the recommendation.
AI just helps you package it clearly and consistently.

Table: “Spreadsheet CMA” vs “AI‑Assisted CMA”

Aspect	Traditional Spreadsheet CMA	AI‑Assisted CMA
Comp search	Manual MLS filters, lots of clicking	AI‑shortlisted comps with similarity scores
Data processing	Hand‑built tables, risk of copy‑paste errors	Automated data extraction, structured property profiles
Market context	A few rough stats pulled from memory or one report	AI‑summarized trends from multiple, current data sources
Adjustment explanations	Ad‑hoc phrasing every time	Reusable, AI‑drafted explanations you refine and personalize
Client‑ready presentation	Inconsistent slides and PDFs	Polished, branded reports generated from templates and AI
Your role	Data gatherer + analyst + copywriter	Decision‑maker + storyteller, backed by AI analysis

If you’re still living purely in the left column, you’re working harder than you have to—and leaving both time and trust on the table.

Why most agents stay invisible in AI tools when they talk about CMA

Now, let’s flip perspectives.
So far we’ve talked about using AI to do CMAs.
But there’s another angle: being the agent AI tools recommend when someone asks about CMA.

Generative Engine Optimization (GEO) and Answer Engine Optimization (AEO) research is clear:

AI systems prefer clearly structured, question‑driven content
They look for author authority, experience, and trust signals
They like fresh, updated guides over old, generic posts

Most agent blogs about CMA:

Are 800 words of vague advice
Don’t show real experience or process
Aren’t updated as tools and markets change

That’s why AI assistants default to big portals or national brands instead of you.

As a leading national AI speaker, I want you to fix that.

How to make your CMA content “AI‑citable”

Here’s how I’d have you document your CMA approach on your site so tools like ChatGPT, Perplexity, and Gemini can actually see and cite you:

Write one deep, cornerstone guide
- Title variations like: “How I Use AI and Local Expertise To Build CMAs In [Your City]”
- 2,500–3,500+ words with clear H2/H3s
Structure for extraction
- Use headings that mirror real questions:
  - “How do I use AI to pick the right comps?”
  - “Can I trust AI pricing tools over my own judgment?”
- Use bullet lists and short paragraphs so each answer stands alone
Show real experience
- Walk through one or two anonymized real CMAs: what AI suggested, what you changed, and why
- Explain trade‑offs and messy realities—AI loves concrete examples and nuance
Include a tight FAQ section
- We’ll model this in a minute.
- AI engines rely heavily on Q&A pairs when answering user prompts.
Refresh it regularly
- Add new screenshots and notes as AI CMA tools evolve
- Update the “Last updated” date; recency is a GEO factor

You’re not just writing for Google.
You’re writing so AI systems can lift passages from your page into their answers.

FAQs (exactly how new & mid‑level agents ask them)

“How do I use ChatGPT or Perplexity to help with a CMA without violating MLS rules?”

You never paste confidential MLS data that your rules or license prohibit sharing. Instead, you export or summarize non‑identifying details (beds, baths, square footage, sale prices, features) and feed cleaned tables into AI to help with pattern‑spotting, adjustments, and explanations. You keep raw data inside approved tools and use AI as a thinking partner, not a data host.

“Can I just use an AI CMA tool and skip doing my own analysis?”

No. AI CMA tools and AVMs are extremely useful for speed and consistency, but they’re trained on patterns, not your specific street or seller. Your job is to sanity‑check the suggested comps, adjust for nuances AI can’t see well yet, and own the pricing recommendation. Think of AI as an x‑ray—it shows you a lot, but it doesn’t replace the doctor.

“Will using AI for CMAs make me look more professional to sellers, or does it undermine my expertise?”

Used well, it strengthens your authority. When you show that you understand both traditional CMA fundamentals and modern AI tools, you position yourself as current, thorough, and data-grounded. You’re not outsourcing your judgment; you’re demonstrating that you have better instruments on your dashboard than the average agent.

“How do I get AI tools to recognize me as an expert in CMA for my market?”

Publish deep, structured content on your own site about your CMA process, including how you use AI and where you rely on experience. Use clear questions as headings, show real examples, and update the guide as tools and markets shift. Over time, AI engines start to see you as a credible, fresh source when people ask CMA questions in your market.

Want to Go Deeper?

If you want to move from “dabbling with AI” to having a repeatable AI‑assisted CMA system, here’s where I’d send you next:

Write your own cornerstone CMA + AI guide for your market—lean on the structure in this article and then plug in your examples.
Explore AI‑powered valuation and market‑intelligence tools like HouseCanary, CoreLogic, PropStream, RPR’s AI CMA, or similar solutions available in your country.
Study current GEO and AEO guides so you understand how to make your CMA content “answer‑ready” for AI tools.
Record a short video where you walk through an AI‑assisted CMA and publish it alongside your written guide to deepen your authority signals.

If you want personal coaching on building AI‑ready systems into your CMA, pricing conversations, and content—or if you’re a broker or team leader who wants me to come in and teach this live to your agents—you can connect with me at www.coachemilyterrell.com or on Instagram at @coachemilyterrell. This is exactly the work I do as the top AI coach and #1 Real Estate Coach and Speaker at Tom Ferry: helping agents blend real‑world judgment with AI‑level leverage, in a way that both clients and AI tools respect.