How to Build AI Agents That Fix Themselves When Something Goes Wrong ft. Keith Gillespie

This week I'm joined by Keith Gillespie, founder of REI Automated, a Marine Corps veteran who spent eight years on active duty across 13 countries, missing the births of both his kids, and came home with one obsession: building a real estate investing business in 1 to 2 hours a day. That constraint forced him to build SOPs that became systems, systems that became software, and software that 5 years ago became REI Automated, a three-pillar platform combining education, CRM, and coaching for investors under 50 deals done on their own.

What started as a personal solution has grown into a fully commercialized system running 24 AI agents, pushing 24 to 70 software updates per day, and built on the philosophy that education alone gets 99% of people nowhere unless the tools and support are right there alongside it. Keith is one of the few people in this industry who has genuinely melded the methodology with the machine, and this conversation gets into exactly how he did it.

Episode Timeline & Highlights

[0:49] – Jordan introduces Keith Gillespie from REI Automated and previews the REI Tech Unlocked event in Dallas and his upcoming AI book

[3:42] – Keith's background as a real estate investor for 10 years across 34 states, and how REI Automated was built to solve his own problem

[4:22] – What active duty Marines life actually looked like: 13 countries in 8 years, missing both kids' pregnancies, and needing to build a business in 1 to 2 hours a day

[5:47] – The three pillars of REI Automated: education, software, and coaching, and how they work together as a system

[7:08] – How the Marine Corps OODA loop (Observe, Orient, Decide, Act) transfers directly to real estate negotiation and relationship management

[10:45] – The grit and tenacity piece: why embracing the suck from military training carries into every hard stretch of building a business

[11:12] – Jordan parallels the OODA loop to the flight training acronym DODAR and why slowing down leads to better decisions than reactive action

[13:34] – Why Keith built REI Automated as a true ecosystem rather than a CRM with coaching bolted on, and what being the company he wished existed ten years ago actually means

[14:25] – Keith's 57 completed real estate investing courses and what he learned by going through virtually every major program in the industry

[15:16] – Why education alone fails 99% of the time and why the fishing pole analogy captures everything wrong with most real estate programs

[16:38] – Why all three pillars are non-negotiable and what happens when any single ingredient is pulled from the cake

[20:02] – The two things any SaaS business actually needs to survive: results and customer service, and how Keith built REI Automated around both

[22:39] – Inside the AI agent build pipeline: discovery, internet research, planning, build, testing, human QA, and production, all in about 50 minutes per feature

[25:36] – How REI Automated scaled from 6 AI agents to 24 and why parsing agents to narrow, specific jobs improves quality at every level

[26:27] – The self-healing agent approach: agents required to update their own project instructions when they learn something new

[28:14] – Building a support ticket rewriter agent to fix poorly worded user requests before they reach the development queue

[34:16] – Who REI Automated is built for: any investor who has done fewer than 50 deals independently, from zero to experienced acquisition managers

[35:53] – Why 389 built-in automations and everything under one roof is the real differentiator, not any individual feature

[37:36] – Zero customer churn since public launch in March 2026, and what a nine-person human staff plus 24 AI agents means for responsiveness

5 Key Takeaways

1. Systems Before Scale — Keith built the foundation of REI Automated out of necessity, not ambition. With only 1 to 2 hours per day to work his real estate business while on active duty, every SOP had to be airtight. The constraint forced better engineering than most funded teams ever produce.
2. Education Without Execution Is Worthless — After completing 57 real estate investing courses, Keith's conclusion is clear: knowledge without a system to apply it and support to keep you accountable is just expensive entertainment. All three pillars have to exist together or the whole thing falls apart.
3. The OODA Loop Works Off the Battlefield — The Marine Corps' observe-orient-decide-act decision framework isn't just for room clearing. It's a real-time relationship and negotiation tool that helps you read a situation, adjust, decide, and act in a loop rather than reacting impulsively and blowing the deal.
4. Agent Quality Comes from Narrow Scope — REI Automated's AI infrastructure improved dramatically when they stopped asking agents to do multiple things and started assigning each one a single, tightly defined job. The more focused the scope, the higher the output quality, and the easier it is to diagnose what breaks.
5. The Answer Is Yes — Keith's product development philosophy is built on one rule: if a client requests something that isn't wrong or illegal, build it. Listening to the marketplace and acting on it fast is why REI Automated has had zero churn since public launch, pushing 24 to 70 updates a day.

Links & Resources

• Keith Gillespie on Facebook — search "Keith Gillispie" (note the spelling: G-I-L-L-I-S-P-I-E)
• SmrtPhone — the only phone system built for real estate investors, includes 5,000 minutes free — smrtphone.io
• REI Tech Unlocked event — September 19–21, Dallas, Texas (co-headlined with Left Main, Stephanie Betters, Steve Trang, Objection Proof AI; sponsored by Facebook, Twilio, Salesforce)
• Labor Architecture AI — Jordan's upcoming book on AI and human employees — laborarchitecture.ai
• That Real Estate Tech Guy — thatrealestatetechguy.com

If Keith's approach to systems, AI, and building the company he wished existed ten years ago resonated with you, share this one with an investor in your network who is still duct-taping tools together with Zapier. This conversation is a field report from someone actively building in real time, and those are the ones worth passing around. More high-signal conversations coming next.