New tech always has its detractors. Every new invention, from the car to the TV to the internet, has faced criticism to some degree, and AI is certainly no exception.
It’s hard to know, sometimes, whether using AI in your content will help your real estate brand or send it spiraling into obscurity.
Best-selling author Chris Smith, who has co-founded multiple AI-powered startups, described it as “100% the biggest game changer of our lives, period.” One of Smith’s top uses for AI is in marketing, which is why he’ll be joining us for BAM BBQ this week. If you haven’t registered yet, sign up here.
Today, you get 10 AI marketing ideas from Chris Smith, every one of which has already worked on a real audience before this post existed.
Every idea below names the exact tool Smith used and the steps to replicate it.
Let’s dive in!
AI Tools to Use on Repeat
Every idea below runs through one of a handful of AI tools Smith uses on repeat:
- Prompt Cowboy: a free prompt-writing tool that turns a basic, one-line prompt into a fully structured, expert-level prompt.
- Higgsfield: an AI wrapper that bundles several top image and video models in one place, including Nano Banana Pro.
- Dundy AI (known as Calico AI Luxury Architect before its rebrand): a free custom GPT that turns a single photo into video without any prompting required.
- Claude: used to build lightweight, free tools like a seller net proceeds calculator, no code required.
- NotebookLM: Google’s AI research tool, used here to build a hyperlocal homebuying guide trained on one specific market.
Within Higgsfield, Smith also has access to several other image and video models, including Nano Banana Pro, Nano Banana 2, GPT Image 2, Grok, Google Veo 3, and Seedance 2.
10 AI-Powered Marketing Plays
#1: Fix your lazy prompts before you touch content.
Every AI tool only outputs what it’s given, and most agents type prompts too thin to get anything useful back. Smith fixes that with Prompt Cowboy, free to start with a paid upgrade available.
Here’s the play in three steps:
- Type a normal, basic prompt, something like “create an SEO and AEO optimized neighborhood page for Avalon Park.”
- Prompt Cowboy expands it into a fully structured expert prompt, complete with a defined role and a real content brief.
- Reuse the same fix on listing descriptions, image prompts, and video prompts.
Prompt Cowboy has become one of Smith’s go-to AI tools because it takes a basic prompt and makes it dramatically better, which improves the final output.
View this post on Instagram
#2: Turn Google Earth into a hyperlocal content machine.
Buyers and sellers both respond to proof that an agent knows a neighborhood cold, not just generic praise about “the area.” Smith turns that proof into a single graphic using Google Earth and Nano Banana Pro.
Step 1: Start in Google Earth, zoom into a single neighborhood, switch to Clean mode to strip out labels, flip to 3D, and take a screenshot. From there, pick one play and stick to it:
- Pre-Market Tease, to trigger buyer FOMO and seller curiosity
- Track Record Proof, to show repetition and dominance without bragging
- Amenity Spotlight, to reveal every home in the area with a pool or a specific feature
Step 2: Pick one of three plays and don’t mix them: Pre-Market Tease, Track Record Proof, or Amenity Spotlight.
Step 3: Drop the screenshot into Nano Banana Pro, apply selective highlighting and upscaling, and add one line of text like “I’ve sold 17 homes in this photo.”
Smith even runs the same aerial shot through Google Veo 3 afterward to turn it into drone-quality video.
He built this specifically to turn hyper-local content into inbound DMs.
#3: Screenshot a competitor’s ad and flip it into your own.
Big iBuyers and portals already spend real money getting their ads in front of your exact buyers and sellers. Smith just borrows that reach for himself, building both versions in Nano Banana Pro from an Opendoor ad he screen-shared during his BAM Fest talk:
- A comparison graphic, side by side, showing what a home sells for with an agent versus through Opendoor
- A parody ad, keeping Opendoor’s original creative but swapping in his own line about working with Opendoor meaning waving goodbye to profit
Both work because they use eyeballs the competitor already paid to earn.
#4: Build a co-branded map using famous local landmarks.
People respond faster to a familiar brand name than to an unfamiliar street name. Smith uses that instinct to make a listing feel closer to daily life, building the map in Nano Banana Pro and Higgsfield:
- Place the listing’s address at the center of the map.
- Mark driving times to recognizable local brands, like Starbucks or Disney, instead of generic street names.
The map borrows recognition that already exists instead of building it from scratch.
#5: Design a “we are here” days-on-market graphic for listing presentations.
Every seller starts to worry around the same point in a listing’s life, right when it crosses the local average days on market. Smith built a graphic in Nano Banana Pro that gets ahead of that conversation, and building it takes two steps:
- Find your local average days on market.
- Design a simple horizontal slide showing that average, with a marker for where a given listing sits on the same timeline.
Bring it to a listing presentation to keep a seller calm before they start asking why the home hasn’t sold.
#6: Make “text in the wild” graphics that look 100% real.
Handwritten, unpolished content reads as personal instead of produced, and personal is what gets people to stop scrolling.
Smith builds these in Nano Banana Pro, designing graphics to look like handwritten whiteboard notes, smudges included, and lets them pass as genuine social content.
He estimates that out of 100 people, about 95 would assume the image is real, and he says these graphics get more interaction than almost anything else he posts.
#7: Turn a viral concept into a coffee cup quote graphic.
A quote sitting inside an everyday object gets more attention than the same quote posted on its own. Smith borrowed this idea after seeing Gary Vee post a Starbucks cup with handwriting on it, then built his own version in Nano Banana Pro. The execution is just two steps:
- Take a photo of a coffee cup.
- Add a quote to it, styled to look like it was written by hand.
He calls it one of his most viral Instagram posts so far.
#8: Turn one listing photo into a cinematic video with zero prompting.
Video still gets more reach than static photos almost everywhere. Most agents just don’t have the time or the editing skill to shoot and produce it themselves. Smith found a shortcut that turns a photo he already has into video with almost no effort, using a free custom GPT called Dundy AI (the rebrand for Calico AI Luxury Architect Image to Video):
Here’s the entire workflow in four steps:
- Screenshot any listing photo and save it.
- Upload it to the GPT, available in the GPT store.
- Skip the prompt entirely. The GPT reads the image itself and writes its own prompt.
- Let it generate the video.
He ran this on a still aerial shot of his own neighborhood and turned it into moving, drone-style footage.
#9: Build your own free lead-magnet tool instead of another blog post.
Sellers care about exactly one number when they’re deciding whether to sell: what they’ll actually walk away with. Smith built a free tool, using Claude, that answers that question instantly instead of writing another blog post about it. Here’s how he built and deployed it:
- Build a simple calculator as a Claude Artifact, in this case a seller net proceeds estimator.
- Give away the tool with no lead capture required.
- Pair it with a script agents can send directly to sellers, offering to estimate cash at close after payoff, commission, and closing costs.
Smith has found that a free tool consistently beats a blog post or a market update in getting people to click.
#10: Build an AI homebuying coach trained only on your market.
Buyers are already asking ChatGPT and Gemini their home buying questions before they ever call an agent, and Zillow already built a national AI guide to capture that traffic. Smith found a way to make sure a buyer’s AI assistant hands them the agent’s name instead of a competitor’s, using Google’s NotebookLM.
The build breaks down into five steps:
- Create a notebook named something like “Buying a Home in [City]: The AI Guide.”
- Load 20 to 40 hyper-local sources:
- Neighborhood guides
- MLS market reports
- Down payment assistance pages
- Local closing process explainers
- County-specific tax and insurance pages
- Add a shorter “unfair local stack” that a national portal genuinely can’t compete with, including county property tax assessor pages and local lender rate sheets.
- Generate the studio outputs: an audio overview, an interactive deep dive, a mind map, and a set of slides, all from the same source stack.
- Upload one more document titled “Meet [Name]: Your [City] Home Buying Expert,” with a process breakdown, specialty neighborhoods, and a few testimonials.
Once that’s in the notebook, asking it how to pick an agent surfaces the agent’s own name instead of a competitor’s, turning the whole notebook into an always-on, top-of-funnel lead magnet.
Nobody needs to run all 10 of these by Friday. Pick two or three that fit tools already sitting on a phone or laptop, ChatGPT, Claude, or a free custom GPT, and build just one this week.
Smith is bringing more of this directly to BAM BBQ, live on July 22. Spots are filling fast, so save yours while you still can.
