← Milee · Related topics · FAQ
AI product support for e-shops: an agent that knows your catalog
Milee loads your product XML feed (Heureka, Zbozi.cz and Google) and turns it into a knowledge base the AI voice agent uses to answer customer questions about specific products on the line – availability, price and specs. For Shoptet it works as an add-on: the feed and data are already in your shop. Plus abandoned-cart recovery, product recommendations and order status.
Key figures
~70%
average cart abandonment rate across studies
Source: Baymard Institute
72%
customer satisfaction with voice AI (up from 53% in 2022)
Source: Voice AI market reviews 2026
-40%
reduction in average call handle time
Source: Voice AI market reviews 2026
$80B
expected labor savings from conversational AI (2026)
Source: Gartner
Best practices
- Connect your product feed (Heureka/Zbozi/Google) – Milee builds a knowledge base from it.
- The agent answers only from feed data (price, availability, specs) – no making things up.
- Call abandoned carts quickly and offer help or a discount.
- Recommend by parameters (buying guidance) and offer accessories.
- Sync the feed regularly so prices and stock stay correct.
FAQ
How does the agent know my products?
From your product XML feed (Heureka, Zbozi.cz or Google). Milee loads it, builds a knowledge base and the agent answers availability, price and specs from it.
Does it work with Shoptet?
Yes. Shoptet generates feeds natively, so you just connect the feed URL; deeper integration (cart, orders, customers) runs as an add-on via the Shoptet API.
Can the agent invent a price or availability?
No – it answers only from feed data. When it lacks the information it says so and offers an email link or a handoff to a human.
Sources
- Baymard Institute – Cart Abandonment Rate Statistics
- Heureka – XML feed specification
- Shoptet Developers – API
- Gartner – $80B call center savings (via CRM Curator)
Related topics
- Speed-to-lead: why the first 5 minutes decide
- Reactivating dormant leads: revive a database you already paid for
- AI voice agent: what it is, how it works and when it pays off
- Reducing no-shows: appointment reminders that work
- Measuring satisfaction (NPS): why phone beats email
- AI vs human call center: how much you really save
- Lead qualification: fast, cheap and without losses
- AI in customer support: what to automate and how much to save
- Follow-up strategy: how many times and when to call
- GDPR and AI calling: what to watch out for
Updated 2026-07-24.