<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Graft AI</title>
    <link>https://graft.axcelner.com/blog</link>
    <description>Writing about dependable agent access, software interfaces, and verified action in systems of record.</description>
    <language>en</language>
    <lastBuildDate>Wed, 15 Jul 2026 00:00:00 GMT</lastBuildDate>
    <atom:link href="https://graft.axcelner.com/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Your Agent Is Ready. Your Business Software Isn’t.</title>
      <link>https://graft.axcelner.com/blog/agents-need-contracts</link>
      <guid isPermaLink="true">https://graft.axcelner.com/blog/agents-need-contracts</guid>
      <pubDate>Wed, 15 Jul 2026 00:00:00 GMT</pubDate>
      <description>Why capable AI agents still need a stable contract for the business software where real work happens—and why Graft starts with the interface.</description>
    </item>
    <item>
      <title>A Click Is Not Proof: Verifying AI Agent Actions</title>
      <link>https://graft.axcelner.com/blog/a-click-is-not-proof</link>
      <guid isPermaLink="true">https://graft.axcelner.com/blog/a-click-is-not-proof</guid>
      <pubDate>Wed, 15 Jul 2026 00:00:00 GMT</pubDate>
      <description>Why reliable agent action requires source-system evidence, explicit permissions, and failure behavior—not just a completed click path.</description>
    </item>
    <item>
      <title>The Workflow Access Gap in Enterprise AI</title>
      <link>https://graft.axcelner.com/blog/the-workflow-access-gap</link>
      <guid isPermaLink="true">https://graft.axcelner.com/blog/the-workflow-access-gap</guid>
      <pubDate>Tue, 14 Jul 2026 00:00:00 GMT</pubDate>
      <description>Why enterprise AI agents stall when critical actions exist only inside legacy or internal interfaces—and how to evaluate the integration options.</description>
    </item>
  </channel>
</rss>
