Yes. SleekAI's data sources read _wc_cog_cost on product posts, _wc_cog_item_cost on order items, and _order_total_cost plus _order_total_profit on shop orders. Variations and variable products use their own per-variation cost meta, which is also exposed. Setup takes a few minutes per data source.

No, as long as you scope the bot correctly. Display conditions in SleekAI let you restrict by user role and URL, so the COGS bot only runs when the visitor is a shop manager or admin on a back-office page. A separate sales bot for the storefront has no access to cost variables in its system message.

Yes. SleekAI reads order data through the WooCommerce data layer, so legacy post-based orders and High-Performance Order Storage both work. Cost of Goods writes its line item cost meta the same way under both storage modes, and the bot resolves the values at request time.

Margin is computed at request time using the order line total minus tax and shipping minus the line item cost. The plugin already exposes _order_total_cost and _order_total_profit per order, so for store-wide aggregates the bot can answer without recomputing anything. The system message describes how to handle gross versus net.

Yes. Either prefilter the data source by a date window (for example, last 30 days of orders) or let the bot call a function that returns filtered totals. Common date-bound questions like worst margins this month or best products last quarter resolve cleanly with a date-restricted data source.

Each variation is a child post with its own _wc_cog_cost. SleekAI's data sources can read the parent and the variations together, so the bot can answer whether the medium-blue variant has a worse margin than the small-red one without you exposing the entire SKU list to the prompt.

Yes, if you wire it up. SleekAI exposes a function-call layer where the bot can trigger PHP actions, like setting a new _regular_price or moving a product to a different category. Most teams keep this read-only to start and add write actions for one or two well-tested workflows.

Conversation logging is per chatbot and stored inside WordPress with messages, model name, token usage, and originating page URL. You can see what kinds of margin questions repeat, fold them into the standup template, or simplify the system message to answer them more directly.