<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0">
  <channel>
    <title>NestedBit Insights</title>
    <link>https://nestedbit.com/insights</link>
    <description>Practical thinking for operating with AI.</description>
    <language>en</language>
    <lastBuildDate>Sun, 30 Aug 2026 00:00:00 +0530</lastBuildDate>
    <item>
      <title>When to build an AI agent</title>
      <link>https://nestedbit.com/insights/when-to-build-an-ai-agent</link>
      <guid>https://nestedbit.com/insights/when-to-build-an-ai-agent</guid>
      <pubDate>Sun, 30 Aug 2026 00:00:00 +0530</pubDate>
      <description>Use the Rule of R to decide whether a workflow deserves an agent.</description>
    </item>
    <item>
      <title>AI agent vs chatbot: the operational difference</title>
      <link>https://nestedbit.com/insights/ai-agent-vs-chatbot</link>
      <guid>https://nestedbit.com/insights/ai-agent-vs-chatbot</guid>
      <pubDate>Sun, 30 Aug 2026 00:00:00 +0530</pubDate>
      <description>A chatbot helps with a conversation. An agent participates in a workflow.</description>
    </item>
    <item>
      <title>Human approval is a system boundary</title>
      <link>https://nestedbit.com/insights/human-approval-for-ai-agents</link>
      <guid>https://nestedbit.com/insights/human-approval-for-ai-agents</guid>
      <pubDate>Sun, 30 Aug 2026 00:00:00 +0530</pubDate>
      <description>Approval gates make consequential AI actions inspectable and accountable.</description>
    </item>
  </channel>
</rss>
