How OpenAI o1 might revolutionise AI for Real Estate Professionals

Did you feel the earth move again?

This time, it’s not a random weather event but another major shift in the AI world—and it’s not about virtual tours or the latest facts and figures.

You may have heard a bit lately about Open AI’s o1 model (often referred to as “Strawberry”) which represents yet another significant leap forward in AI capabilities that we can all use to supercharge our daily lives.

“How is it different from what we have now?”

As my primary job is still “writer,” I’m going to use a writing analogy to try to explain this simply.

A standard Large Language Model (like ChatGPT, or Claude, or Gemini will behave like this:

You throw a prompt at them, and off they go, writing their response one sentence at a time, just like a first draft or someone thinking out loud (albeit usually better than my first drafts!!).

It’s the acting equivalent of improv or a TV show like “Thank God you’re here!”

Basically, the model just keeps talking until they decide they are done—like a salesperson without empathy or appropriate listening skills.

But, just like in real-life situations, this can be where things can start to go sideways.

Since your average large language model reasons on the fly, they can easily contradict themselves in the same breath they made the most emphatic statement you’ve ever read.

One minute, they confidently tell you one thing, and the next, they’ve decided the opposite might also be true.

“Take a deep breath and think step-by-step…”

Remember this?

A study released in February this year analysed the commands that elicited the best responses from ChatGPT.

The results showed that prompts such as "take deep breaths" and "think step by step" ranked highly, suggesting that these phrases can help improve the quality of the AI's responses.

It is meant to change the behaviour of the AI by getting it to figuratively take a deep breath and walk you through its thinking instead of blurting out the first thing that pops into its electrons.

But o1 is different.

I recall my math teacher's advice during exams: it wasn't just about getting the right answer, but also showing how you got there.

If your answer was incorrect, but your step-by-step process was sound, you'd still earn some credit.

This approach taught me the value of working through a process in problem-solving – a principle that's equally relevant when working with AI models like o1.

The difference between OpenAI o1 and other versions of the model

I hope I am not oversimplifying when I say if someone suddenly shoved a mic in your face and demanded a quick answer, you might ramble a bit. That’s ChatGPT 3.5/4.

But if they gave you a few minutes to gather your thoughts, jot some notes, and polish your response, you’d come across way more put together. That’s OpenAI-o1.

That’s basically what chain-of-thought prompting does for the AI. It gives it a chance to “get its act together” before putting its keystrokes into gear.

To be clear, o1-preview doesn’t do everything better. It actually might not be a better writer than GPT-4o, for example. 

But for tasks that require planning, the differences are significant.

So if you happen to be thinking: “Do I really need to care about the difference between one AI model and another? I’m a real estate agent, not a tech expert.”

You know I am going to say the answer is “yes”, so if you can, stick with me because this latest leap in AI isn't just some behind-the-scenes upgrade—it can directly impact how we run our businesses, handle clients, and even close deals.

Let's break it down and explore how the differences in models might apply to real estate.

Traditional LLMs: Great at Language, But Limited in Thought?

The large language models we’ve been using so far—like GPT-3 or GPT-4—are amazing at understanding and generating human-like text.

But these models have a key limitation: they’re not really “thinking” or reasoning.

They’re pulling from a huge database of patterns in language to guess what words come next in a sentence.

That’s why they’re great at tasks like:

• Writing property descriptions

• Summarising market trends

• Answering common client questions

• Creating templates for emails, social posts, or newsletters

However, they struggle when you ask them to handle more complex, dynamic tasks—like helping you work out what a $400,000 mortgage at 6 percent interest for 30 years might cost if interest rates are adjusted upwards of .5 percent each month for the next 6 months.

Language with Reasoning

While previous models excelled at generating text, the o1 model introduces something more powerful: reasoning.

This means it’s not just predicting the next word in a sentence based on what it’s seen before—it can analyse information, draw conclusions, and make decisions based on more complex relationships between data points.

Here’s a basic example to illustrate the difference:

• Old LLM: If you ask it for the best time to sell a property, it might give you a broad, generic answer based on seasonal trends—like, “Spring is typically a great time to sell.”

• O1 Model: Ask the same question, and it could analyse your specific market conditions, look at comparable listings, assess local stock levels, and even consider external factors like upcoming zoning changes or interest rate forecasts. It might come back with, “Given current market conditions in your area and the recent uptick in interest rates, it would be wise to list in the next x weeks, as buyer demand will likely drop after that.”

For a ‘real life’ comparison, o1 reasoned it’s way through the mortgage problem I listed above pretty well.

How the o1 Model Might help agents in the future

So, what does this reasoning ability mean for you guys in real estate?

Of course, the possibilities are enormous.

Here are a few thoughts on where the o1 model could be - dare I say it - a game-changer:

1. Property Valuation and Investment Analysis

Imagine working with an investor who’s considering multiple properties. Previously, ChatGPT might generate some basic comparisons—square footage, location, price, etc.

With o1, the AI can go deeper with its increased mathematical superpowers.

It could probably consider each property's investment potential based on historical trends, current market conditions, and your client’s goals. It may even simulate different scenarios, like how a new development nearby might impact future property values. This might allow you to offer clients a more comprehensive and personalised evaluation of their investment options, taking into account a broader range of factors and potential outcomes.

2. Client Personalisation

Do you keep track of a client’s journey through the property buying or selling process in your CRM?

Using its reasoning superpowers, models like o1 might be able to analyse data from all your client interactions, such as properties viewed, lead forms filled out, texts sent, emails sent … and so on.

By analysing this data, models like o1 can identify patterns and subtle preferences that may not be immediately apparent, such as a client's consistent prioritisation of having a home office or proximity to top-rated schools.

This would allow listing agents and buyers agents to refine their property suggestions and provide more personalised recommendations, without having to manually read between the lines. It might even allow you to suggest homes with the same or similar characteristics in more budget friendly areas.

3. Contract and Deal Negotiation

When it comes to negotiations, your human instincts and skills as an agent are always going to be your paycheck.

And while no AI can (or should) replace this, models like o1 model might give you an edge by helping you weigh various factors—like buyer urgency, competing offers, and potential future value—faster and more accurately.

It might even suggest negotiation strategies and role plays based on past outcomes in similar situations, helping you tailor your approach to each unique deal.

4. Market Predictions and Trend Analysis

If you’ve ever tried to predict where the market is going, you know it’s a bit like trying to see around a corner without x-ray vision.

Traditional LLMs can summarise trends, but they don’t “reason” through data in the same way the o1 model probably could.

The o1 model might take extra steps and look at economic reports, housing inventory, interest rates, and other variables, and offer reasoned predictions about future market shifts. Which, in turn, would give you the opportunity offer clients informed advice about whether to buy, sell, or hold onto a property.

***

Hopefully, by now, you can see that the o1 model isn’t just a “better” version of existing language models.

It’s fundamentally different and represents a new capability that could once again, change how you work—saving you time, improving client service, and offering insights that previously would’ve required a team of economists and mathematicians.

At its core, the o1 model blends the best of what LLMs already do—handling language tasks effortlessly—with an ability to reason and problem-solve. So while the old models were like having a super-efficient assistant, the o1 model is more like having an advisor who can help you think through decisions.

For real estate agents - and the rest of us - that opens up a whole new world of possibility.

If you have a paid version of ChatGPT you can take a bite of that strawberry, test it out and think of it as another smart way to stay ahead in a business that’s always evolving.

Happy Hunting 🍓