Feed Quality Product Data

Your AI chatbot is only as smart as the data you give it. A well-structured product feed is the foundation of accurate, helpful recommendations.

Why Product Data Quality Matters

When a customer asks "I'm looking for red running shoes under 100 EUR in size 42," the AI needs to search through your product catalog. If your data is missing colors, sizes, or accurate prices, the chatbot can't deliver what the customer needs — no matter how smart the AI is.

Lobot uses hybrid search — combining semantic understanding (what the customer means) with keyword matching and metadata filtering. The richer your data, the more powerful this search becomes.

What Makes a Great Product Feed

• Detailed product names — "Nike Air Max 90 Men's Running Shoe - White/Black" beats "Nike Shoe"
• Rich descriptions — Include materials, use cases, key features, and benefits. The AI uses these to understand what the product is really about.
• Accurate pricing — Always keep prices current. Stale prices destroy customer trust.
• High-quality images — Lobot displays product cards with images. Blurry or missing images hurt conversion.
• Proper categorization — Well-organized categories help the AI narrow down results faster.
• Size, color, and variant data — Essential for filtering. Lobot automatically deduplicates size variants of the same product.

Supported Feed Formats

Lobot supports the most common e-commerce feed formats and auto-detects the structure:

• Heureka XML — standard for Czech and Slovak e-shops
• Izsoft XML — popular in the Czech e-commerce ecosystem
• Shopify API — direct integration with Shopify stores
• Custom XML — any well-structured XML feed

Simply provide the feed URL in your admin panel — Lobot handles the rest, including automatic synchronization at your chosen interval.

How Automatic Sync Works

Once configured, Lobot automatically:

1. Fetches your feed at the set interval (e.g., every 60 minutes)
2. Detects changes — only updated products are re-processed
3. Converts HTML descriptions to clean text
4. Generates vector embeddings for semantic search
5. Updates the product database without any downtime

Common Data Quality Issues

• Missing attributes — No color or size data means the bot can't filter effectively
• Outdated prices — Products showing wrong prices erode customer trust
• Broken image URLs — Product cards without images look unprofessional
• Duplicate products — Lobot handles size variants automatically, but true duplicates should be cleaned in your feed

Key Takeaways

• Invest time in rich, detailed product descriptions — it directly improves AI recommendations
• Keep your feed data current — enable automatic sync
• Include all relevant attributes: name, description, price, images, category, color, size
• Lobot auto-detects your feed format — just provide the URL