The RealHomes property edit screen is famous for the sheer number of meta boxes it ships with. Basic Information, Additional Details, Energy Performance, Floor Plans, Virtual Tour, Agent details, and a Field Builder for everything else. Buyers do not page through all of that. They just type their question into a chat box and expect a straight answer.

SleekAI sits on top of the RealHomes data model. The bot reads the property custom post type, the REAL_HOMES_property_price, REAL_HOMES_property_bedrooms, REAL_HOMES_property_bathrooms, and REAL_HOMES_property_size postmeta, the property-type, property-status, property-city, and property-feature taxonomies, and the linked agent profile so it can quote price, beds, baths, and the listing agent without ever rendering a page.

Bolt a generic chatbot onto a RealHomes site and it answers from whatever happened to be visible above the fold. Tabs, accordions, and Field Builder values all get ignored. SleekAI pulls live values on each turn, so when a property moves from Active to Reserved the chatbot already knows by the next question.

RealHomes is a deeply customizable theme. The property post type holds the basics, but the real power lives in postmeta. The REAL_HOMES_ keys store price, beds, baths, garage, and energy data.

The Field Builder lets agencies add anything else that matters in their market: council tax band, freehold or leasehold status, ground rent, service charge, broadband speed, EPC rating. Buyers ask about exactly those fields. Generic chatbots that read whatever HTML happens to render do not see most of that data.

They miss the Field Builder values entirely. They miss the EPC if it sits inside an accordion. They quote a price that has not been updated since last week's render and have no idea whether the listing has been reserved.

A SleekAI bot is built directly on the property CPT and the postmeta layer. When a visitor asks about EPC rating, it pulls the value from the exact meta key you mapped. When they ask whether the listing is still available, it reads property-status taxonomy and returns the answer that matches the current database state.

When they want to compare two listings, it pulls both posts at once and returns a side-by-side answer that quotes price, sqft, and EPC for each. The outcome on the brokerage side is the boring kind of good. Fewer wasted agent emails about already-sold properties.

More booked viewings on the listings that are still active. Cleaner inquiry logs because every chat ends with either a clear next step or an agent transcript already in the right inbox.