Yes. WPForms stores entries in wpforms_entries and field values in wpforms_entry_fields, both of which SleekAI can be configured to read as data sources. The bot's system message exposes form metadata and recent entries as variables, so the model has structured data to answer with rather than a wall of HTML.

Not unless you configure it that way. Display conditions can lock the entry-reading bot to administrators or specific custom roles, and SleekAI also supports user-capability checks. Any public bot you run on the same site is a separate chatbot under multibot, with its own data sources and no access to entries.

Yes. Multibot lets you run several chatbots on one install, each with its own system message, model, data sources, and display conditions. A typical pattern is one public sales bot scoped to product pages and one admin bot scoped to the dashboard, with no overlap in what either can read.

Push older or archived entries into an OpenAI Files vector store of up to one gigabyte per file and let the assistant retrieve only the entries that match a query. Recent entries can stay live in the data source so date-range questions still feel real-time, while history goes into retrieval.

Yes. The bot can read core entry fields regardless of which WPForms tier you are on, as long as entry storage is enabled. Some surfacing helpers, like form-level summaries, depend on fields that exist in both Lite and Pro, so the basic flow still works on the free version.

Yes, but make it a different bot. Run a customer-facing chatbot that only knows about public form definitions and your knowledge base, separate from the admin entry-reading bot. Each has its own data sources and display conditions, so the public bot has no path to entry rows.

If the form captures UTM or referrer fields, yes. WPForms can store hidden fields, and SleekAI reads whatever ends up in wpforms_entry_fields. The bot can roll entries up by source, give counts per channel, and compare against a previous period when you ask.

Calls to entry data happen at the moment a question is asked, not on every page load, and SleekAI uses scoped queries rather than dumping the whole table into the prompt. For very large entries archives, retrieval through OpenAI Files keeps both response time and token usage bounded.