<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Djordje Puzic</title>
    <description>The latest articles on DEV Community by Djordje Puzic (@vectorintel).</description>
    <link>https://dev.to/vectorintel</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4065618%2Fd4bb41a9-4732-4550-8bd9-20714192e8ce.png</url>
      <title>DEV Community: Djordje Puzic</title>
      <link>https://dev.to/vectorintel</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/vectorintel"/>
    <language>en</language>
    <item>
      <title>Version control for decisions: why AI-native agencies need change-control</title>
      <dc:creator>Djordje Puzic</dc:creator>
      <pubDate>Thu, 06 Aug 2026 09:42:35 +0000</pubDate>
      <link>https://dev.to/vectorintel/version-control-for-decisions-why-ai-native-agencies-need-change-control-cah</link>
      <guid>https://dev.to/vectorintel/version-control-for-decisions-why-ai-native-agencies-need-change-control-cah</guid>
      <description>&lt;p&gt;AI coding agents are getting better every month. Most agency delivery failures I see are not model failures.&lt;/p&gt;

&lt;p&gt;They are &lt;strong&gt;decision drift&lt;/strong&gt;.&lt;/p&gt;

&lt;p&gt;A client changes a constraint in Slack. A senior engineer encodes a different assumption in a PR. An agent invents a third version of the truth because every session starts cold. Rework follows.&lt;/p&gt;

&lt;h2&gt;
  
  
  What “change-control” means here
&lt;/h2&gt;

&lt;p&gt;Think less chatbot history, more &lt;strong&gt;version control for decisions&lt;/strong&gt;:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Capture decisions and constraints with provenance (who said what, where, when)&lt;/li&gt;
&lt;li&gt;Detect conflicts before they become rework&lt;/li&gt;
&lt;li&gt;Keep humans as the authority on what becomes project truth&lt;/li&gt;
&lt;li&gt;Give coding agents a live preflight against that truth (MCP helps)&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  A practical stack shape
&lt;/h2&gt;

&lt;p&gt;For 3–15 person shops shipping with Cursor / Claude Code / MCP clients, the useful layer is:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Ingest&lt;/strong&gt; — Slack, GitHub, docs, voice notes into project-scoped memory&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Reconstruct&lt;/strong&gt; — current state with supersedes-aware retrieval&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Arbitrate&lt;/strong&gt; — owners resolve conflicts; only approved updates become truth&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Expose&lt;/strong&gt; — agents query constraints before they act&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Self-host + BYOK matters when client project data cannot sit on a random vendor SaaS.&lt;/p&gt;

&lt;h2&gt;
  
  
  Where I’m putting this
&lt;/h2&gt;

&lt;p&gt;I’m building &lt;strong&gt;Accord Book&lt;/strong&gt; around that shape: self-hosted AI memory and change-control for AI-native agency teams.&lt;/p&gt;

&lt;p&gt;Product page: &lt;a href="https://vector-intelligence.io/accord-book/" rel="noopener noreferrer"&gt;https://vector-intelligence.io/accord-book/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Founding pilot: &lt;a href="https://vector-intelligence.io/pilot/" rel="noopener noreferrer"&gt;https://vector-intelligence.io/pilot/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;If you’re an agency owner wearing tech-lead + PM + client-translator hats at once, I’d love to hear where decision drift hits you hardest.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>productivity</category>
      <category>mcp</category>
      <category>agents</category>
    </item>
  </channel>
</rss>
