Chapter 5: Learning to Listen

From answering questions to understanding what buyers actually need.

For a long time, real estate websites have been built around the same basic interaction:

Search. Filter. Click.

You enter a price.

Choose a location.

Select a number of bedrooms.

Scroll through the results.

It works.

But it doesn't really tell you much about the person doing the searching.

And that started to bother me.

Because when someone says:

“Show me homes under $900,000.”

That's not really the whole question.

Why are they moving?

Do they need to be close to work?

Do they have kids?

Do schools matter?

Do they need a garage?

Are they willing to renovate?

Do they want to be close to transit because they don't drive?

Are they looking for a home to live in, an investment, or both?

The more I thought about the concierge experience, the more obvious it became:

A good concierge doesn't just answer the question you're asked. They listen for the question behind it.

So I started thinking about what that could look like on a real estate website.

Instead of forcing someone to communicate entirely through filters, what if they could simply talk?

They could say what they're looking for in their own words.

The system could ask questions.

Clarify what matters.

Remember the context.

And gradually build a better understanding of what that person is actually trying to accomplish.

That became the foundation for the AI Real Estate Concierge.

But the goal wasn't to bolt a chatbot onto a real estate website.

There are already plenty of those.

The goal was to make conversation another way of navigating the entire platform.

A buyer could ask about a property.

Then ask about the neighbourhood.

Then compare it with another home.

Then ask about schools nearby.

Then explain that they need to commute downtown without a car.

The conversation could evolve naturally because the system had context.

The AI wasn't just responding to the latest message.

It was starting to understand the person behind the search.

That changed how I thought about the entire platform.

The MLS provides the inventory.

The map provides the geography.

Liked Listings provide memory.

Saved Searches provide continuity.

And now conversation provides context.

Each layer makes the next one more useful.

That was the moment the project stopped feeling like a collection of real estate features.

It started feeling like a concierge.

Something that could listen, understand, remember, and eventually help.

And that opened up a much bigger question:

If the system can understand what someone wants, what else can we do with that information?

That question would lead directly into the next stage of the project.

Turning conversations into intelligence.

Chapter 6: Understanding People.