Yes. FunnelKit stores funnels as wfacp_funnels posts and steps as wfacp_main posts with _wfacp_options postmeta, and orders carry a _wfacp_funnel_id key on the shop_order. SleekAI can be configured to read all of those via the data source wizard, so the bot reasons over real rows instead of guessing from rendered pages.

Only if you configure it that way. Display conditions can lock the analytics bot to administrators or a custom growth role. A separate public bot under multibot can run on the same site with no access to shop_order tables, so revenue figures stay scoped to the right WordPress users only.

Yes. FunnelKit records bump products on the order through line items and meta, and step posts hold the configured bump products. The bot can count orders that accepted the bump versus declined, group by funnel and step, and return a take rate percentage instead of a raw list of orders.

Push historical orders into an OpenAI Files vector store and let retrieval pull only the relevant slice on demand. Recent orders stay live in the data source so date-range queries still feel real-time, while last quarter's funnel history lives in vector retrieval for trend questions.

Yes. The bot can read core wfacp_funnels and wfacp_main rows on the free version. Pro-only features like global checkout, A/B testing, and one-click upsells add extra meta keys that SleekAI can read once you enable Pro, but the basic funnel and order picture works without it.

Yes. SleekAI ships a PHP embed plus a JS API so you can launch the bot from any hook or admin notice. A common pattern is a Funnel Lookup button on the FunnelKit Funnels list screen that opens the chat scoped to whichever funnel ID the admin just clicked.

Yes. The bot can group orders by created date and return week-over-week or month-over-month deltas per funnel and per step. That makes it easy to ask whether the new checkout layout actually lifted conversion after launch, without leaving WP Admin to build a custom report.

Bring your own API key for OpenAI, Anthropic, Google, or OpenRouter. Each chatbot in the multibot setup can pick a different model, so a public help bot can run on a cheap model while the admin analytics bot uses a stronger reasoning model for the FunnelKit reporting work.