Tidio’s Product Recommendations is an exciting AI innovation, where Lyro recommends products to your visitors when they ask for specific product advice.
This guide focuses on making Product Recommendations possible by connecting your product feed directly to Lyro through a file (which must be compatible with Google Merchant Center formatting). If you are using Shopify or WooCommerce, we have dedicated integration methods that will allow you to use Product Recommendations through their dedicated systems; please see our detailed guides on that method here (for Shopify) and here (for WooCommerce). In addition to that, please see this article to learn about the Product Recommendations engine via OpenAPI, if that is a method that suits you better.
In this article, you will learn:
- Who Product Recommendations (via Product Feed upload) is available to
- How Product Recommendations work in practice
- What conversation types can be handled
- How to set up Product Recommendations in Tidio
- How to pause Product Recommendations
- What the limitations are
Who can use Product Recommendations (via Product Feed upload)?
Product Recommendations lets your Lyro AI agent assist your visitors with product questions. For Lyro to answer these questions, you will need to connect your product database with your Tidio project - which is done by directing Lyro to your raw product feed file. This will allow your product database to be utilized as knowledge for Lyro. Please note that Lyro understands your product information in multiple different languages!
As long as you have a correctly-formatted file, you can connect it with Lyro, and that will serve as the product feed for the Product Recommendation engine. This method is platform-agnostic, meaning that it will work for any store, as long as it is done properly. Importantly, the file must be compatible with Google Merchant Center, and in XML format.
The product feed URL must be completely public; not protected by any password, VPN, or IP filtering. If you are using any security plugins on your website site, you might also need to ensure automated fetchers aren't getting blocked.
To use Product Recommendations, you also need to be a subscriber of the Lyro AI Agent package in Tidio (or the custom Plus package, of course). The relevant developer resources will become available for use on your Tidio project then.
Please note that the Product Recommendations tool is not available during the initial free trial.
Please note: Lyro can detect a visitor's current URL automatically, which allows you to use this information in product-related responses. This is especially valuable if Lyro is asked about a product that the visitor is currently viewing on your website. Lyro will be aware of the product being browsed and won't need to ask the visitor about the details first, nor use any previously-recorded data.
Product Recommendations in practice
Once you synchronise Lyro with your product database (we cover this later in the article), your AI agent will be able to guide your customers to the right products. Lyro will be proactive, asking questions that help find what the customer is looking for, and offering options and advice wherever appropriate.
Any recommendation Lyro has will be presented in the form of cards, which can be scrolled horizontally inside the Tidio widget. The cards include pictures, product names, and prices. Each card can be clicked to visit the product page directly as well, in addition to the View details link.
This helps your customers find products and make decisions quickly, without having to browse your store manually in search of the perfect item. If you're using Shopify, please note that the View details link is replaced by the direct Add to cart button.
Conversation types handled by Product Recommendations
Using Product Recommendations, Lyro can help your visitors in the following ways:
- Specific product search - find a specific product
- Broad product recommendation - help choose products
- Product variant recommendation - find products with variant limitation
- Complementary product recommendation - find products that work well together (if mentioned in product description)
- Alternative product recommendation - find products with similar properties
- Product comparison - compare two (or more) products
- Price range recommendation - find products based on price point
- Seasonal product recommendation - find new products or products in collections
Setting up Product Recommendations
To begin your setup, navigate to the Lyro AI Agent section in your Tidio panel, and open the Products tab:
Inside, use the Sync products button to see the available options:
Inside, you will see the available options: WooCommerce product sync (which is described separately in this guide), Shopify product sync (described in this guide), and Add products via API (described in this guide). The option we're covering here is Product feed upload:
Once you click on that option, a popup will appear, prompting you for the file address:
Please note: you must provide the direct link to the file itself, and make sure that the URL is publicly accessible.
Once the upload finishes, your product database will become visible in the Products tab, along with some basic details (and the option to decide whether the database should be used by Lyro and/or Copilot).
From now on, Lyro (and Copilot) will be able to connect with your product database and refer to your products during conversations.
Pausing Product Recommendations
You can temporarily disable Product Recommendations anytime, from within the Products tab inside your Lyro AI Agent section. Notice the Used by options at the bottom, where you can disable (or enable) the Lyro and Copilot switches:
This allows you to keep Lyro (and/or Copilot) from using your Product Recommendations database. You can switch any of these options back on again whenever you like.
Product Recommendations limitations
Lyro AI Agent subscription
Using Product Recommendations is only possible if your project is subscribed to the Lyro AI Agent package.
Visitor currencies
Only one currency is supported per product feed upload, and it is detected during your first import only. If the currency in your target file is changed, the product feed would need to be uploaded again.
Comments
0 comments
Please sign in to leave a comment.