<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Blog — AI chatbots in practice | SmartChatBot</title><description>Writing about AI chatbots in practice: what works, what does not, and what we learned building for real clients.</description><link>https://smartchatbot.rs/</link><item><title>AI hallucination: why &quot;we trained it well&quot; is not an answer</title><link>https://smartchatbot.rs/en/blog/ai-hallucination-why-a-better-prompt-is-not-the-fix/</link><guid isPermaLink="true">https://smartchatbot.rs/en/blog/ai-hallucination-why-a-better-prompt-is-not-the-fix/</guid><description>Every LLM bot invents a fact sometimes — that is a property of the technology, not a defect. What separates vendors is whether they solved it structurally or are hoping it will not come up.</description><pubDate>Wed, 15 Jul 2026 00:00:00 GMT</pubDate><category>From practice</category><category>Trust</category><author>Marko Marković</author></item><item><title>What customers actually ask a bot (and what you think they ask)</title><link>https://smartchatbot.rs/en/blog/what-customers-actually-ask-a-bot/</link><guid isPermaLink="true">https://smartchatbot.rs/en/blog/what-customers-actually-ask-a-bot/</guid><description>Two findings from production that change how you read your own bot&apos;s data: why your list of &quot;top questions&quot; is probably measuring your own interface, and why customers ask for your company number before they ask your price.</description><pubDate>Wed, 15 Jul 2026 00:00:00 GMT</pubDate><category>From practice</category><category>Knowledge base</category><author>Marko Marković</author></item></channel></rss>