SleekAI logs every conversation in WordPress and can trigger webhooks or emails to your CRM. Most agencies forward qualified leads with the full transcript attached, so the agent gets context immediately rather than starting from scratch. Reapit, Alto, Vebra, Acquaint, and custom CRMs all work via webhook, with payload mapping configurable per integration.

Yes. If you use ACF, Meta Box, or custom post meta for fields like EPC rating, council tax band, freehold/leasehold status, or service charge, SleekAI reads those into the prompt as named variables. Whatever your theme stores on the listing post, the bot can answer about. Hidden internal fields like commission rate stay excluded by default.

Mark the listing as sold or unpublished and SleekAI will reflect that immediately. It only quotes from currently published, available listings, so the bot will not surface a property that is already under offer or completed. For under-offer status, you can use a custom field that the bot mentions but the listing remains visible for comparable property queries.

Yes. Multibot lets you scope each chatbot to its own region or office, with different greeting messages, listing scopes, and even different agent personalities. A central London office and a regional branch can run distinct chatbots with no cross-contamination, each routing to the appropriate team. Per-bot field selection lets each scope expose only the fields relevant to its market.

No. Using your own OpenAI API key with the standard API terms means conversation data is not used for model training by default. That keeps proprietary information about your listings, leads, and qualification flow private to your account. For agencies handling sensitive transactions, this is the correct architecture, and OpenAI's enterprise terms strengthen that further if needed.

SleekAI uses OpenAI Files as a vector store for big knowledge bases, so even agencies with thousands of listings get fast, accurate answers. Only the most relevant 5 to 20 listings are pulled into each conversation based on the visitor's query, which keeps token costs predictable. As listings change, the vector store updates without manual rebuilds.

If your WordPress site is the source of truth for listings, the bot reads your live posts. If listings come from a third-party feed, the bot still reads whatever is published on your WordPress pages. For agencies that maintain a separate Rightmove or Zoopla feed and only mirror it to WordPress, the chatbot answers from the WordPress mirror, so freshness depends on your feed sync schedule.

Lettings work the same way. The bot reads each lettings listing's monthly rent, deposit, available-from date, and tenancy terms, and qualifies tenants on income, references, pet status, and move-in date. Many agencies run separate sales and lettings chatbots via multibot, since the qualification questions differ enough that splitting the configurations makes sense.