The chatbot only recommends what it receives from the product feed. If your descriptions are sparse, brief or lack key attributes, the chatbot won’t be able to match the product to the customer’s needs. In this article, we’ll show you how to prepare product data to ensure the best possible recommendations.
Why does the quality of the feed matter?
When a customer writes “I’m looking for a gift for a 10-year-old for up to 100 zł”, the chatbot must analyse your products and select those that match this description. It can only do this if the feed contains:
information about the target group (children, adults, age)
price
product category and use
Without this data, the chatbot will give a vague response or suggest unsuitable products.
What should a good product description include?
Essential elements
Product name – precise, including type and variant
❌ "Sweatshirt 123"
✅ "Men’s fleece jumper – dark green, size M"
Product description – at least 3–5 sentences that answer the following questions:
What is it for?
Who is the target audience?
What problem does it solve or what need does it meet?
How does it stand out from similar products?
Price – always up to date, synchronised with the feed
Attributes – the most important element for recommendations
The more attributes, the more precise the match. Depending on the industry, it is worth considering:
Industry | Key attributes |
Fashion | Size, colour, material, cut, gender, season |
Electronics | Brand, model, compatibility, technical specifications |
Home and garden | Dimensions, material, style, use, room |
Cosmetics | Skin type, ingredients, effect, capacity |
Children and toys | Age, safety, material, play category |
Availability
Make sure the feed is updated regularly. The chatbot should not recommend products that are out of stock – this frustrates customers and undermines trust in the tool.
How to improve your product feed?
Go to the settings of your shop platform (Shopify, WooCommerce, PrestaShop, etc.)
Check which fields are filled in for each product
Fill in any missing attributes and descriptions
Synchronise the feed with edrone
Tip: Start with the bestsellers category – improving the descriptions there will yield the quickest results.
What else can you do besides the feed?
Add shopping guides to the chat knowledge base that link products to needs:
"How do I choose the right bike size?" → the chatbot will learn to ask the right questions and filter products
"What mattress do we recommend for someone with back pain?" → the chatbot will use this knowledge when dealing with health-related queries
"Round-up: 5 best gifts under £20" → a ready-made sales script for the chatbot
Common mistakes in product descriptions
Copying descriptions from the manufacturer – often generic, without tailoring them to your customer
Lack of keywords describing the use – the chatbot won’t ‘guess’ that someone is looking for hiking boots for a mountain trail if you don’t specify this
Out-of-date prices or stock levels – the chatbot recommends a product that isn’t in stock