Yes. SleekAI reads the standard product custom post types and post meta from both WooCommerce and EDD, including variations, downloads, bundles, and subscription products. For stores with custom CPTs - say, a separate 'wholesale' product type - those are picked up too. The integration doesn't require a separate sync or export step; product data flows from WordPress directly to the prompt context at request time.

Stock is read from post meta on each request, so the chatbot reflects whatever the WooCommerce inventory shows at that moment. If a shopper buys the last unit while another shopper is mid-conversation, the second conversation will reflect the change on the next message. For stores with separate POS-driven stock systems, the bot sees whatever WooCommerce currently has - which is usually the source of truth or close to it.

It can confirm policies, ETAs, and shipping windows from your published content. For specific order lookups by number, you can wire it to a custom endpoint with display conditions for logged-in users only - so a customer asking 'where's order #48211' triggers the bot to fetch their order data and answer with the actual carrier and tracking. Anonymous users see policy answers; logged-in customers get personal answers.

The widget loads asynchronously and only fetches the full context payload when a user actively starts chatting. There's no upfront catalog dump on page load and no synchronous request blocking the page render. Lighthouse scores are typically unaffected. For high-traffic stores, the bot's compute happens at the moment of intent rather than per pageview, which keeps the cost predictable and the page snappy.

All conversation logs stay in your WordPress database. Nothing is sent to Sleek servers - the data path is browser-to-WordPress-to-OpenAI directly via your API key. This matters for stores with strict data-residency requirements, GDPR-conscious EU buyers, or contractual obligations that limit third-party data sharing. Set retention to match your privacy policy and use the standard WordPress database export for audit and DSAR workflows.

Yes. Multibot scopes a bot to product categories, tags, specific SKUs, or custom display-condition rules. A women's-apparel bot with sizing guidance can run on those categories; a tech-accessories bot with compatibility checks can run on others. Shared content - shipping, returns, brand voice - lives in a common knowledge pool. For stores with 5,000+ SKUs across distinct departments, this is the architecture that keeps answers focused and accurate.

Yes - configure the system prompt to surface related or complementary products from the same category when a shopper has expressed clear interest. Light suggestions ('the Kestrel 38L pairs well with the Trail rain cover, also in olive') work better than aggressive cross-sell scripts. The bot reads category and tag relationships from your catalog, so suggestions track your real merchandising rather than a separate recommendation engine.

Add your shipping policy, customs notes, and duty-and-tax content to an OpenAI Files vector store and the bot quotes the relevant paragraph for the shopper's destination. A shopper in Canada asking about duties on a 200 EUR order gets the Canada-specific note rather than the generic policy. For stores shipping to many countries, this is materially better than a single shipping FAQ that buries the country detail.