Sharing Mike DelPrete experiment because it shows, better than any slide, what I’ve been telling portals for years.
the interface was never the moat. The data is.
I've been a buyer, a seller, a broker, an office manager and a franchise owner.
In 22 years I never once chose a portal because of its interface.
Nobody I know did either.
That's why Mike's experiment hit me.
When an AI assistant reads listings directly and already knows your life, the search screen stops mattering. Anyone can generate one. Thousands of them. What nobody can generate is the inventory and what you know about it.
Mike talks about a second chance for MLSs. In Europe we don't have an MLS. The portal is the inventory. So this is a second chance for portals. If they take it.
Here is the uncomfortable part. "Just open the data" is a trap. A portal that only feeds AI becomes a utility. Utilities don't set prices. Someone else does.
So what keeps a portal relevant?
Clean data, open APIs, plugins. That's the entry ticket, not the strategy.
Data the AI can't get elsewhere. The assistant knows the user. The portal knows the market: verified supply, price history, what sold, which broker delivers. That's a trade. Business models for sharing with AI will get built. The question is by whom.
And the user. Years ago I built a tool that did one thing: it told a homeowner what their home was worth. The portals had nicer websites. 450,000 people registered with us anyway. They came for the number.
Waze grew the same way. It competed with Google successfully because it gave drivers what they needed: live data from other drivers. Not a prettier map.
The ones who already go to AI do it for exactly this reason. The portal doesn't help them compare and judge which offers fit them. That is one of the biggest gaps in the portal business. And it's not a product gap. It's a business model gap. What's valuable to a buyer or a seller is often what the broker doesn't like. And the broker pays.
Portals solved this once already. They put the buyer first while the broker paid for the listing. That choice is the trust they live on today. Doing it again won't be easy. It's still the only move.
One more thing on interfaces. If they're infinite, don't try to design the winning one. Let users build their own. Let them share the best ones. Reward the ones whose interface others adopt. Host the market, don't fight in it.
Bring the brokers into the same loop and you get what every portal says it wants: a safe, easy place to buy and sell, and leads that are worth something.
Portals have always held two things: the listings and the search. AI is now trying to take the search. Today it's maybe one search in a hundred. The number isn't the point, the direction is. Interesting to see how this plays out. What are your thoughts?