How We Teach AI

AI is changing real estate. Unsurprisingly, that means a lot of people want to learn how to use it.

As a place that teaches AI, this has been great for us at Thesis Driven. But it's not without complications. Given the speed at which AI is transforming real estate, it needs to be taught thoughtfully, with a focus on real student outcomes and practical applications, not hype and generalizations.

As we gear up to teach our latest AI in Real Estate workshop tomorrow, I wanted to share a bit about how we approach the most transformative technology of our lifetimes.

Imagine you're teaching a cooking class. The following people are enrolled:

• A young bachelor who wants to learn how to cook pasta.
• A pasta chef looking to discover the latest trends in Italian cuisine.
• A restaurateur scaling from 3 to 10 locations, looking to hire pasta chefs.
• A restaurant critic who believes everyone is overcooking their pasta.

Now imagine they all believe their livelihoods depend on getting really, really good at cooking pasta over the next 12–24 months lest they be cast into a permanent low-carb underclass.

Our students land in the classroom at every stage of their AI journey, each with their own goals and constraints. I've been here before: I co-founded General Assembly in 2010, which expanded to more than twenty brick-and-mortar campuses and 35,000 students learning technology and design skills.

To teach AI effectively and pragmatically, we lean on a few principles that serve us and our students well:

1. Stay Practical
The biggest mistake we see people make when teaching AI is going too theoretical. Given the diversity of students and goals that inevitably end up in any open program, it's tempting to keep things high level. But that's an error; students enroll in programs like ours because they want tactical insights, not academic theory.

Our rule of thumb: students who take our programs should be able to cover the workshop cost at least 10x over with applications they can implement the very next day.

2. Segment by Goal, Not Role
The level of pragmatism described in that last rule can only be achieved by an outcomes-oriented approach. "What is your job title?" is far less important than "What do you want to achieve?" Figure that out, and the curriculum flows from there.

This mitigates the "cooking class" problem illustrated earlier. If someone is looking to hire chefs, they simply shouldn't be in the class.

In other words, teach to the outcome.

3. Lean on Thought Experiments
Our "AI in Property Management" workshop is actually titled Building the Zero Employee Property Manager. The entire frame of the workshop is how someone would hypothetically build a property management firm with zero (human) employees.

We don't do that because it would be a good idea to do so. In fact, we'd strongly discourage it.

We teach AI in Property Management through that example because it forces students to consider the sticky parts of implementation: automation, data silos, PMS integration, handoffs, and risk. When there are no humans in the mix, those things matter far more.

Otherwise, it's too easy to get stuck walking through point solutions vendor-by-vendor.

4. Ground it in industry Realities
There are a million "Learn AI in 12 minutes" videos on YouTube. People take programs with us at Thesis Driven because we are native to the real estate industry and understand its quirks and limitations.

There are plenty of great generalist AI tools out there. Some of them have exceptional applications within the real estate industry. Others are a poor fit. It's impossible to know the difference without understanding how the industry works: operations, capital markets, development, and construction.

5. Give the Vendor List
Everyone wants vendor lists. They want to see our take on the top vendors in AI leasing, land search, maintenance triage, fraud detection, and whatever else we touch in class.

We resisted this at first. AI is moving so fast, any vendor list we supply today is going to be outdated in six months. It felt much more useful to focus on frameworks and evaluations so operators in our programs could vet any vendor claim rather than just relying on whatever we tell them.

But providing the vendor list is part of promise to stay practical. Students want to apply the learnings the very next day. So we made the lists. We teach the frameworks in the course and give the vendor lists out after the program.

We're thrilled to keep offering more programming on AI; we're rolling out new workshops almost every week. You can see the full list here and consider checking out tomorrow's AI in Real Estate workshop!

-Brad Hargreaves