<?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>Loy O'Kelley — Insights</title>
    <link>https://loyokelley.com/insights</link>
    <atom:link href="https://loyokelley.com/feed.xml" rel="self" type="application/rss+xml" />
    <description>Field notes on running enterprise programs with AI in the loop — governance, reporting, program risk, and the delivery stack.</description>
    <language>en-us</language>
    <copyright>Copyright 2026 Loy O'Kelley</copyright>
    <managingEditor>loy@loyokelley.com (Loy O'Kelley)</managingEditor>
    <webMaster>loy@loyokelley.com (Loy O'Kelley)</webMaster>
    <lastBuildDate>Wed, 26 Aug 2026 13:00:00 +0000</lastBuildDate>
    <image>
      <url>https://loyokelley.com/apple-touch-icon.png</url>
      <title>Loy O'Kelley — Insights</title>
      <link>https://loyokelley.com/insights</link>
    </image>
    <item>
      <title>The Machine That Cried Wolf: AI Alert Fatigue on Programs</title>
      <link>https://loyokelley.com/the-machine-that-cried-wolf</link>
      <guid isPermaLink="true">https://loyokelley.com/the-machine-that-cried-wolf</guid>
      <pubDate>Wed, 26 Aug 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>An early-warning system that flags everything warns you of nothing. Alert fatigue is the failure mode of AI risk detection — and triage is the governance work nobody scoped.</description>
    </item>
    <item>
      <title>The Recorder in the Room: Every Meeting Is a Record Now</title>
      <link>https://loyokelley.com/recorder-in-the-room</link>
      <guid isPermaLink="true">https://loyokelley.com/recorder-in-the-room</guid>
      <pubDate>Wed, 19 Aug 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>Nobody approved turning your steering committee into a verbatim transcript. It happened anyway. What that costs, and how to govern it.</description>
    </item>
    <item>
      <title>Subcontracted Judgment: Your Vendors Are Delivering With AI Too</title>
      <link>https://loyokelley.com/subcontracted-judgment</link>
      <guid isPermaLink="true">https://loyokelley.com/subcontracted-judgment</guid>
      <pubDate>Wed, 12 Aug 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>Your AI policy stops at your badge line. Your accountability doesn&#x27;t. What to ask vendors, and what to put in the contract.</description>
    </item>
    <item>
      <title>Chain of Custody: Where Did That Number Come From?</title>
      <link>https://loyokelley.com/chain-of-custody</link>
      <guid isPermaLink="true">https://loyokelley.com/chain-of-custody</guid>
      <pubDate>Wed, 05 Aug 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>AI-drafted artifacts arrive without a paper trail. Traceability is the control that decides whether your program can defend its own numbers.</description>
    </item>
    <item>
      <title>Judgment Atrophy: AI Is Taking the Reps That Made Your Team Good</title>
      <link>https://loyokelley.com/judgment-atrophy</link>
      <guid isPermaLink="true">https://loyokelley.com/judgment-atrophy</guid>
      <pubDate>Wed, 29 Jul 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>AI takes the routine work first — and the routine work is where judgment was built. How to keep training the bench you&#x27;ll need when the automation is wrong.</description>
    </item>
    <item>
      <title>The Confidence Problem: AI Sounds the Same When It&#x27;s Wrong</title>
      <link>https://loyokelley.com/the-confidence-problem</link>
      <guid isPermaLink="true">https://loyokelley.com/the-confidence-problem</guid>
      <pubDate>Wed, 22 Jul 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>A human hedges when they&#x27;re unsure — and that hedge is information. AI strips it out. How to engineer doubt back into AI-augmented program reporting.</description>
    </item>
    <item>
      <title>Mind the Seams: Where AI-Augmented Programs Actually Break</title>
      <link>https://loyokelley.com/the-seams</link>
      <guid isPermaLink="true">https://loyokelley.com/the-seams</guid>
      <pubDate>Wed, 15 Jul 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>AI made the inside of every team fast and visible. Programs still break at the seams between teams — the handoffs and dependencies no one owns.</description>
    </item>
    <item>
      <title>Automation Debt: Every Automation You Build Is a Liability You Now Own</title>
      <link>https://loyokelley.com/automation-debt</link>
      <guid isPermaLink="true">https://loyokelley.com/automation-debt</guid>
      <pubDate>Wed, 08 Jul 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>The build is the cheap part. The upkeep is the program. Why every automation you stand up is a standing liability someone has to own.</description>
    </item>
    <item>
      <title>Pilot Purgatory: Why AI Programs Die After the Demo</title>
      <link>https://loyokelley.com/pilot-purgatory</link>
      <guid isPermaLink="true">https://loyokelley.com/pilot-purgatory</guid>
      <pubDate>Wed, 01 Jul 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>The demo is the easy 10%. Production is the job. Why AI pilots stall after proof of concept — and how to build them to ship.</description>
    </item>
    <item>
      <title>The Rubber-Stamp Review: When the Human in the Loop Stops Looking</title>
      <link>https://loyokelley.com/rubber-stamp-review</link>
      <guid isPermaLink="true">https://loyokelley.com/rubber-stamp-review</guid>
      <pubDate>Wed, 24 Jun 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>When AI is usually right, the reviewer stops reading — and the safeguard goes hollow. How to keep human oversight a real control.</description>
    </item>
    <item>
      <title>The Decision Backlog: Programs Stall on Decisions, Not Work</title>
      <link>https://loyokelley.com/decision-backlog</link>
      <guid isPermaLink="true">https://loyokelley.com/decision-backlog</guid>
      <pubDate>Wed, 17 Jun 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>The real bottleneck on most programs isn&#x27;t the work — it&#x27;s the decisions waiting on someone. Make decision latency visible before it costs you the schedule.</description>
    </item>
    <item>
      <title>Shadow AI Is Already on Your Program</title>
      <link>https://loyokelley.com/shadow-ai-on-your-program</link>
      <guid isPermaLink="true">https://loyokelley.com/shadow-ai-on-your-program</guid>
      <pubDate>Sun, 14 Jun 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>A ban doesn&#x27;t stop AI on your program — it just removes your ability to govern it. Bring it into the light.</description>
    </item>
    <item>
      <title>When the Agent Does the Work, Who Signs?</title>
      <link>https://loyokelley.com/ai-accountability</link>
      <guid isPermaLink="true">https://loyokelley.com/ai-accountability</guid>
      <pubDate>Wed, 10 Jun 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>The accountability boundary — the one design decision that decides whether AI makes your program safer or just faster.</description>
    </item>
    <item>
      <title>Garbage In, Gospel Out: AI Inherits Your Data Discipline</title>
      <link>https://loyokelley.com/garbage-in-gospel-out</link>
      <guid isPermaLink="true">https://loyokelley.com/garbage-in-gospel-out</guid>
      <pubDate>Wed, 10 Jun 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>AI doesn&#x27;t fix a dishonest system of record — it launders it into confident, polished fiction. Data hygiene is the precondition for everything else.</description>
    </item>
    <item>
      <title>The AI-Augmented PMO: What Actually Changes</title>
      <link>https://loyokelley.com/ai-augmented-pmo</link>
      <guid isPermaLink="true">https://loyokelley.com/ai-augmented-pmo</guid>
      <pubDate>Sun, 31 May 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>What AI and agentic systems really change inside a program management office — and what they don&#x27;t.</description>
    </item>
    <item>
      <title>The Delivery Stack Behind an AI-Augmented Program</title>
      <link>https://loyokelley.com/ai-delivery-stack</link>
      <guid isPermaLink="true">https://loyokelley.com/ai-delivery-stack</guid>
      <pubDate>Sun, 31 May 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>Claude, Copilot, Power Automate, Make, and n8n — the real tools and the governance layer behind AI-augmented delivery.</description>
    </item>
    <item>
      <title>AI Won&#x27;t Save a Failing Program — But It Warns You Sooner</title>
      <link>https://loyokelley.com/ai-program-risk</link>
      <guid isPermaLink="true">https://loyokelley.com/ai-program-risk</guid>
      <pubDate>Sun, 31 May 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>AI for early risk detection in distressed programs — what it catches, and what still needs a human.</description>
    </item>
    <item>
      <title>Killing the Status-Report Tax with AI</title>
      <link>https://loyokelley.com/status-report-tax</link>
      <guid isPermaLink="true">https://loyokelley.com/status-report-tax</guid>
      <pubDate>Sun, 31 May 2026 13:00:00 +0000</pubDate>
      <category>AI-Augmented Delivery</category>
      <dc:creator>Loy O'Kelley</dc:creator>
      <description>How to use AI to automate project reporting — and why the goal is honesty, not just speed.</description>
    </item>
  </channel>
</rss>
