Almost every week, a new artificial intelligence tool promises to put real estate agents ahead. And almost every week, Phil Stringer watches agents pay for one, barely use it and move on to the next.
“One of the big roadblocks right now for real estate agents specifically is that they’re trying to do too much because there’s so many shiny objects,” Stringer told Inman.
Stringer is founder and CEO of The AI Academy and calls himself an AI efficiency coach. In his first year at General Electric, he was its top sales agent worldwide. He later became COO of one of North Carolina’s top real estate brokerages, according to his website. He has used AI since before ChatGPT went mainstream, starting with data analysis and early writing tools like Jasper.
Inman spoke with Stringer last month at eXpCon, eXp Realty’s convention in Salt Lake City. Here’s his advice on getting more out of AI.
Pick one tool and stick with it
Agents already wear a lot of hats, Stringer said, and learning new software every month is the last thing they need.
“If you try and do every single thing that’s new, you will constantly be switching directions every single week,” he said. “It’s just not good for your business.”
The same goes for large language models like ChatGPT, Gemini, Claude, Perplexity and Grok. Stringer said he hates being asked which one is best, because his answer could change within a week. He compares it to choosing a CRM: The best one is the one you’ll actually use.
Stringer uses Claude as his daily driver. For things Claude can’t do well, like generating images, he uses other tools. He said ChatGPT is “killer” right now, and he called ChatGPT Images 2.5, which OpenAI released in September, the best image model available.
Still, his core advice holds. “Pick a tool, stick with the tool, don’t jump ship every couple weeks,” he said.
Build a ‘2nd brain’
The most important thing anyone can do with AI today, Stringer said, is build what he calls a “second brain.”
It’s a folder on your computer full of files about your business, and the AI can build it for you. Stringer’s pulls from his text messages, recorded phone calls, email, computer files and CRM data. It also takes in-person conversations recorded by Omi, an AI wearable he uses.
Connect that folder to Claude, ChatGPT or another tool, and the AI can read your full business context. It can also update the folder as you work.
The bigger payoff is portability. If a better model comes out, you point it at the same folder instead of starting over. Without that, switching tools can feel like hiring a new employee every day, he said.
“It’s the most incredible thing we’ve ever done with our business,” Stringer said.
Stringer built his second brain with Claude Code. Scheduled routines in Claude pull in new information every day. He tells it to keep only what’s relevant to his business.
“I don’t need the email from Dick’s Sporting Goods in my second brain,” he said.
It also catches conflicting information. When a product’s price changes from $297 to $497, for example, it archives the old price and keeps the new one.
“My brain is updated as of today, and I can plug it into any tool,” he said.
Use AI for quality, not just quantity
Stringer has no patience for low-effort, AI-generated posts. “I cannot stand AI slop content,” he said. If an account posts bad content, he unfollows it.
He believes AI can help agents produce quality content at scale. “It’s going to literally amplify whatever you give it,” he said. Give it weak material and you get more weak material. Give it strong material and you can post at a level you couldn’t reach before.
One tool he relies on is Higgsfield, which bundles dozens of image, video and audio models under one subscription. (Stringer said he has no affiliation with the company.) He connects Higgsfield to Claude, which picks the right model for each request.
For carousel social posts, he generates an image and then has Higgsfield turn it into a short motion graphic. In one, smoke billows from a rocket ship on the first slide, which catches the eye in a scrolling feed. Stringer said his carousels drew 11 million views over the previous 30 days.
AI won’t do all of it, though. Agents still need to set aside time to review what AI researches for them and rewrite scripts so they sound like themselves.
His other tip is to pick 10 accounts you admire and ask your AI tool to find each one’s outliers: the posts that did much better than usual over the past 30, 60 or 90 days. Big accounts get engagement no matter what they post, he said, so the outliers show what actually worked.
“I’m not saying copy,” Stringer said. He said agents should use those outliers as inspiration to “make something similar but do it in your own voice.”