<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Source Grounding on Fondsites</title><link>https://fondsites.com/tags/source-grounding/</link><description>Recent content in Source Grounding on Fondsites</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Wed, 13 May 2026 16:10:13 +0300</lastBuildDate><atom:link href="https://fondsites.com/tags/source-grounding/feed.xml" rel="self" type="application/rss+xml"/><item><title>AI Agent Knowledge Bases: Keeping Delegated Work Grounded</title><link>https://fondsites.com/ai-agents/guidebooks/agent-knowledge-bases-grounding/</link><pubDate>Tue, 12 May 2026 00:00:00 +0000</pubDate><guid>https://fondsites.com/ai-agents/guidebooks/agent-knowledge-bases-grounding/</guid><description>&lt;p&gt;An AI agent can sound grounded even when it is drifting. It may write in the right tone, cite familiar project names, and produce a plan that feels plausible. The weakness only appears when someone asks where the answer came from. Was it using the latest policy or a copied summary from last quarter? Did it retrieve the canonical design note or a discussion thread where the idea was rejected? Did it treat a customer email as evidence, or did it quietly let that email redefine the process?&lt;/p&gt;</description></item></channel></rss>