YITH WooCommerce Compare gives shoppers a clean side-by-side table of titles, prices, images, and attributes. The catch is that a column of specifications only helps when the shopper knows what each row means. A jacket compare row labeled "hydrostatic head: 10,000 mm" is useful to a hiker and noise to everyone else. SleekAI maps the same product titles, prices, attributes, and taxonomies into the chatbot's system message, so the bot can read the comparison the way a knowledgeable salesperson would and translate the spec rows into a recommendation.

Display conditions scope the bot to compare pages, the floating compare widget, or specific category archives, so the conversational layer shows up exactly where the visitor is staring at a table they cannot fully parse. Multibot lets a beginner-friendly assistant run on consumer categories while a deeper, more technical bot runs on B2B or pro lines, each with its own system message and model.

Every conversation logs to WordPress with the model name, token usage, and the page URL it ran on. Use the JS API to open the bot from the compare table's "Need help choosing?" link with the SKUs pre-loaded, so a hesitant shopper never has to retype what they are comparing.

The compare table is the moment a shopper is closest to buying and most likely to bounce. They have narrowed down from dozens to two or three, but the rows in front of them are full of numbers and units they do not entirely understand. A 3-layer fabric, a hydrostatic head of 10,000 mm, a fill power of 700, those mean something to someone who already knows the category and very little to a first-time buyer.

The default outcome is that the shopper picks the cheapest of the three and treats it as a hedge, or worse, leaves to search for a comparison article on a third-party site and never comes back. A generic chatbot does not help here because it has no idea which three products the visitor is comparing, what your actual prices are, or which attributes you stock. It will invent rows or default to a brand it has seen on the web.

SleekAI reads the exact same WooCommerce attributes that the YITH table renders, and a system message that says "recommend one of these three based on the shopper's use case" turns a wall of specifications into a single sentence of advice. The table is the data, the chatbot is the salesperson who reads the table out loud and points at the right one. Conversations log to WordPress, so the recurring three-way comparisons are easy to spot and either become curated pages or feed back into clearer category copy.