<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
<title>Agile Insights Briefing</title>
<link>https://podcast.agile-insights.com.au/#briefings</link>
<atom:link href="https://podcast.agile-insights.com.au/feeds/briefing.xml" rel="self" type="application/rss+xml"/>
<language>en-au</language>
<copyright>© Agile Insights</copyright>
<description>A current question in AI strategy and governance, discussed from the sources: what happened, what is confirmed, and what an organisation should do about it. From Agile Insights. Drawn from public sources and checked before we publish. Voices are synthetic; scripts are drafted with AI and reviewed by a person.</description>
<itunes:summary>A current question in AI strategy and governance, discussed from the sources: what happened, what is confirmed, and what an organisation should do about it. From Agile Insights. Drawn from public sources and checked before we publish. Voices are synthetic; scripts are drafted with AI and reviewed by a person.</itunes:summary>
<itunes:author>Agile Insights</itunes:author>
<itunes:owner><itunes:name>Agile Insights</itunes:name></itunes:owner>
<itunes:image href="https://podcast.agile-insights.com.au/feeds/briefing.jpg"/>
<image><url>https://podcast.agile-insights.com.au/feeds/briefing.jpg</url><title>Agile Insights Briefing</title><link>https://podcast.agile-insights.com.au/#briefings</link></image>
<itunes:category text="Business"><itunes:category text="Management"/></itunes:category>
<itunes:category text="Technology"/>
<itunes:explicit>false</itunes:explicit>
<itunes:type>episodic</itunes:type>

<item>
<title>Episode 38: Budgeted for seats, billed for work · Part 2: Reading the five times headline</title>
<link>https://podcast.agile-insights.com.au/companies-are-spending-5x-more-on-ai-now/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/companies-are-spending-5x-more-on-ai-now/discussion-natural.mp3</guid>
<pubDate>Thu, 19 Nov 2026 06:38:01 +1100</pubDate>
<description>Menlo Ventures says the average company spent 4.7 times as much on AI as last year, against a plan of 1.9 times. Agents turned AI from a licence per person into a meter that runs on work, so the forecast is built on the wrong unit.

Fact check: 22 claims checked (18 verified, 1 partly right, 3 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/companies-are-spending-5x-more-on-ai-now/</description>
<enclosure url="https://podcast.agile-insights.com.au/companies-are-spending-5x-more-on-ai-now/discussion-natural.mp3" length="4775586" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/companies-are-spending-5x-more-on-ai-now/cover.jpg?v=6e27bdbc"/>
<category>agents</category><category>leadership</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 38: Budgeted for seats, billed for work · Part 1: How AI is priced now</title>
<link>https://podcast.agile-insights.com.au/companies-are-spending-5x-more-on-ai-now/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/companies-are-spending-5x-more-on-ai-now/primer-natural.mp3</guid>
<pubDate>Thu, 19 Nov 2026 06:38:00 +1100</pubDate>
<description>Menlo Ventures says the average company spent 4.7 times as much on AI as last year, against a plan of 1.9 times. Agents turned AI from a licence per person into a meter that runs on work, so the forecast is built on the wrong unit.

Fact check: 22 claims checked (18 verified, 1 partly right, 3 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/companies-are-spending-5x-more-on-ai-now/</description>
<enclosure url="https://podcast.agile-insights.com.au/companies-are-spending-5x-more-on-ai-now/primer-natural.mp3" length="3843643" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/companies-are-spending-5x-more-on-ai-now/cover.jpg?v=6e27bdbc"/>
<category>agents</category><category>leadership</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 37: What is an AI skill (and why you should care)</title>
<link>https://podcast.agile-insights.com.au/2610-custom-gpts-gems-retired-skills-migration/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-custom-gpts-gems-retired-skills-migration/primer-natural.mp3</guid>
<pubDate>Tue, 17 Nov 2026 06:37:00 +1100</pubDate>
<description>A skill is a written recipe an AI assistant picks up by itself when the work matches. ChatGPT, Gemini, Copilot and Claude now share the format, and custom GPTs and Gems are being converted into skills, so each one needs an owner, a home and a test.

Fact check: 15 claims checked (11 verified, 3 partly right, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-custom-gpts-gems-retired-skills-migration/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-custom-gpts-gems-retired-skills-migration/primer-natural.mp3" length="4698861" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-custom-gpts-gems-retired-skills-migration/cover.jpg?v=ec37f9f9"/>
<category>agents</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 36: Checking agents in production, even when you can’t read their work · Part 2: When you can’t read the traces</title>
<link>https://podcast.agile-insights.com.au/2611-legora-agent-evals-without-customer-traces/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2611-legora-agent-evals-without-customer-traces/discussion-natural.mp3</guid>
<pubDate>Mon, 16 Nov 2026 06:36:01 +1100</pubDate>
<description>An agent in production can get worse without anyone touching it. Microsoft’s Foundry guide sets out a running evaluation loop, and Legora shows how to keep that loop running when privacy stops you reading the traces.

Fact check: 11 claims checked (8 verified, 1 partly right, 2 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2611-legora-agent-evals-without-customer-traces/</description>
<enclosure url="https://podcast.agile-insights.com.au/2611-legora-agent-evals-without-customer-traces/discussion-natural.mp3" length="5814971" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2611-legora-agent-evals-without-customer-traces/cover.jpg?v=17b32f62"/>
<category>agents</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 36: Checking agents in production, even when you can’t read their work · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/2611-legora-agent-evals-without-customer-traces/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2611-legora-agent-evals-without-customer-traces/primer-natural.mp3</guid>
<pubDate>Mon, 16 Nov 2026 06:36:00 +1100</pubDate>
<description>An agent in production can get worse without anyone touching it. Microsoft’s Foundry guide sets out a running evaluation loop, and Legora shows how to keep that loop running when privacy stops you reading the traces.

Fact check: 11 claims checked (8 verified, 1 partly right, 2 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2611-legora-agent-evals-without-customer-traces/</description>
<enclosure url="https://podcast.agile-insights.com.au/2611-legora-agent-evals-without-customer-traces/primer-natural.mp3" length="4941001" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2611-legora-agent-evals-without-customer-traces/cover.jpg?v=17b32f62"/>
<category>agents</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 35: Who watches the agents? Detect, stop, and let a person approve the rules</title>
<link>https://podcast.agile-insights.com.au/who-watches-the-agents/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/who-watches-the-agents/audio.mp3</guid>
<pubDate>Sat, 14 Nov 2026 06:35:00 +1100</pubDate>
<description>Google tested its agent monitor on benchmarks it did not write, OpenAI showed how it stops agent jobs, and Perplexity showed how it tunes its security rules with a person signing off. What it means for anyone running agents.

Fact check: 16 claims checked (1 confirmed, 15 reported), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/who-watches-the-agents/</description>
<enclosure url="https://podcast.agile-insights.com.au/who-watches-the-agents/audio.mp3" length="7029769" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/who-watches-the-agents/cover.jpg?v=0487ae31"/>
<category>security</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 34: When the test stops telling you anything: hunger beats tenure · Part 2: Discussion</title>
<link>https://podcast.agile-insights.com.au/ainews-anthropic-ends-coding-interviews-hunger-beats-tenure/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ainews-anthropic-ends-coding-interviews-hunger-beats-tenure/discussion-natural.mp3</guid>
<pubDate>Fri, 13 Nov 2026 06:34:01 +1100</pubDate>
<description>One Anthropic manager has dropped the coding interview because AI now passes it. What that says about every measure built on a skill AI supplies, and about who gets to back an idea.

Fact check: 20 claims checked (14 verified, 2 partly right, 4 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ainews-anthropic-ends-coding-interviews-hunger-beats-tenure/</description>
<enclosure url="https://podcast.agile-insights.com.au/ainews-anthropic-ends-coding-interviews-hunger-beats-tenure/discussion-natural.mp3" length="6427407" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ainews-anthropic-ends-coding-interviews-hunger-beats-tenure/cover.jpg?v=9b94c9df"/>
<category>people</category><category>leadership</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 34: When the test stops telling you anything: hunger beats tenure · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ainews-anthropic-ends-coding-interviews-hunger-beats-tenure/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ainews-anthropic-ends-coding-interviews-hunger-beats-tenure/primer-natural.mp3</guid>
<pubDate>Fri, 13 Nov 2026 06:34:00 +1100</pubDate>
<description>One Anthropic manager has dropped the coding interview because AI now passes it. What that says about every measure built on a skill AI supplies, and about who gets to back an idea.

Fact check: 20 claims checked (14 verified, 2 partly right, 4 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ainews-anthropic-ends-coding-interviews-hunger-beats-tenure/</description>
<enclosure url="https://podcast.agile-insights.com.au/ainews-anthropic-ends-coding-interviews-hunger-beats-tenure/primer-natural.mp3" length="4655381" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ainews-anthropic-ends-coding-interviews-hunger-beats-tenure/cover.jpg?v=9b94c9df"/>
<category>people</category><category>leadership</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 33: Agents in production: the platform runs it, you still own it · Part 2: Discussion</title>
<link>https://podcast.agile-insights.com.au/2610-foundry-multi-agent-production-design-security/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-foundry-multi-agent-production-design-security/discussion-natural.mp3</guid>
<pubDate>Thu, 12 Nov 2026 06:33:01 +1100</pubDate>
<description>A managed agent platform changes where the code runs, not who answers for what the agents do. Every boundary in a multi-agent system needs a named owner and a test that proves it holds.

Fact check: 12 claims checked (10 confirmed, 1 partly right, 1 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-foundry-multi-agent-production-design-security/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-foundry-multi-agent-production-design-security/discussion-natural.mp3" length="5582655" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-foundry-multi-agent-production-design-security/cover.jpg?v=a131b347"/>
<category>agents</category><category>security</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 33: Agents in production: the platform runs it, you still own it · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/2610-foundry-multi-agent-production-design-security/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-foundry-multi-agent-production-design-security/primer-natural.mp3</guid>
<pubDate>Thu, 12 Nov 2026 06:33:00 +1100</pubDate>
<description>A managed agent platform changes where the code runs, not who answers for what the agents do. Every boundary in a multi-agent system needs a named owner and a test that proves it holds.

Fact check: 12 claims checked (10 confirmed, 1 partly right, 1 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-foundry-multi-agent-production-design-security/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-foundry-multi-agent-production-design-security/primer-natural.mp3" length="5368527" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-foundry-multi-agent-production-design-security/cover.jpg?v=a131b347"/>
<category>agents</category><category>security</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 32: Agentic video is real: what to try, and what to check · Part 2: Discussion</title>
<link>https://podcast.agile-insights.com.au/agentic-video-is-real/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/agentic-video-is-real/discussion.mp3</guid>
<pubDate>Tue, 10 Nov 2026 06:32:01 +1100</pubDate>
<description>AI agents now edit footage and animate data, so work video is cheap. A video pipeline is software: give it an owner, a number check, consent for every face and voice, and a date to retire.

Fact check: 14 claims checked (5 confirmed, 7 reported, 1 unverified, 1 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/agentic-video-is-real/</description>
<enclosure url="https://podcast.agile-insights.com.au/agentic-video-is-real/discussion.mp3" length="4810810" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/agentic-video-is-real/cover.jpg?v=a7e3a4cf"/>
<category>oversight</category><category>people</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 32: Agentic video is real: what to try, and what to check · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/agentic-video-is-real/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/agentic-video-is-real/primer.mp3</guid>
<pubDate>Tue, 10 Nov 2026 06:32:00 +1100</pubDate>
<description>AI agents now edit footage and animate data, so work video is cheap. A video pipeline is software: give it an owner, a number check, consent for every face and voice, and a date to retire.

Fact check: 14 claims checked (5 confirmed, 7 reported, 1 unverified, 1 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/agentic-video-is-real/</description>
<enclosure url="https://podcast.agile-insights.com.au/agentic-video-is-real/primer.mp3" length="4513883" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/agentic-video-is-real/cover.jpg?v=a7e3a4cf"/>
<category>oversight</category><category>people</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 31: Speeding up planning decisions with AI: who checks the checker? · Part 2: Discussion</title>
<link>https://podcast.agile-insights.com.au/speeding-up-planning-decisions-with-ai/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/speeding-up-planning-decisions-with-ai/discussion-natural.mp3</guid>
<pubDate>Tue, 10 Nov 2026 06:31:01 +1100</pubDate>
<description>England&#x27;s PlanAssist drafts planning reports and an officer approves every one. What its three headline numbers measure, why a high no-edit rate is a question to test, and what a human sign-off does not exempt you from.

Fact check: 10 claims checked (8 verified, 1 disputed, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/speeding-up-planning-decisions-with-ai/</description>
<enclosure url="https://podcast.agile-insights.com.au/speeding-up-planning-decisions-with-ai/discussion-natural.mp3" length="5683116" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/speeding-up-planning-decisions-with-ai/cover.jpg?v=35676f4b"/>
<category>oversight</category><category>people</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 31: Speeding up planning decisions with AI: who checks the checker? · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/speeding-up-planning-decisions-with-ai/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/speeding-up-planning-decisions-with-ai/primer-natural.mp3</guid>
<pubDate>Tue, 10 Nov 2026 06:31:00 +1100</pubDate>
<description>England&#x27;s PlanAssist drafts planning reports and an officer approves every one. What its three headline numbers measure, why a high no-edit rate is a question to test, and what a human sign-off does not exempt you from.

Fact check: 10 claims checked (8 verified, 1 disputed, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/speeding-up-planning-decisions-with-ai/</description>
<enclosure url="https://podcast.agile-insights.com.au/speeding-up-planning-decisions-with-ai/primer-natural.mp3" length="5138583" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/speeding-up-planning-decisions-with-ai/cover.jpg?v=35676f4b"/>
<category>oversight</category><category>people</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 30: The prompt that shows what your AI actually knows about you (and it’s less than you might think)</title>
<link>https://podcast.agile-insights.com.au/the-prompt-that-shows-what-your-ai-actually-knows-about-you/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/the-prompt-that-shows-what-your-ai-actually-knows-about-you/audio.mp3</guid>
<pubDate>Mon, 09 Nov 2026 06:30:00 +1100</pubDate>
<description>Your assistant’s memory is a short file it writes about you: partial, sometimes wrong and harder to erase than it looks. How to see it, correct it and make it forget, and why at work it is a data store someone has to govern.

Fact check: 13 claims checked (7 verified, 3 partly right, 1 disputed, 2 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/the-prompt-that-shows-what-your-ai-actually-knows-about-you/</description>
<enclosure url="https://podcast.agile-insights.com.au/the-prompt-that-shows-what-your-ai-actually-knows-about-you/audio.mp3" length="4732870" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/the-prompt-that-shows-what-your-ai-actually-knows-about-you/cover.jpg?v=807f8e11"/>
<category>data</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 29: CrowdStrike Uncovers New Agentic AI Campaigns · Part 2: Discussion</title>
<link>https://podcast.agile-insights.com.au/crowdstrike-uncovers-new-agentic-ai-campaigns/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/crowdstrike-uncovers-new-agentic-ai-campaigns/discussion-natural.mp3</guid>
<pubDate>Mon, 09 Nov 2026 06:29:01 +1100</pubDate>
<description>CrowdStrike reports four groups running break-ins with Claude Code, set up the way organisations set up their own agents. Why the model’s guardrails are not your control, and the controls that still hold when an agent moves in minutes.

Fact check: 10 claims checked (5 verified, 2 partly right, 1 disputed, 2 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/crowdstrike-uncovers-new-agentic-ai-campaigns/</description>
<enclosure url="https://podcast.agile-insights.com.au/crowdstrike-uncovers-new-agentic-ai-campaigns/discussion-natural.mp3" length="5555198" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/crowdstrike-uncovers-new-agentic-ai-campaigns/cover.jpg?v=7c2f8d0a"/>
<category>security</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 29: CrowdStrike Uncovers New Agentic AI Campaigns · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/crowdstrike-uncovers-new-agentic-ai-campaigns/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/crowdstrike-uncovers-new-agentic-ai-campaigns/primer-natural.mp3</guid>
<pubDate>Mon, 09 Nov 2026 06:29:00 +1100</pubDate>
<description>CrowdStrike reports four groups running break-ins with Claude Code, set up the way organisations set up their own agents. Why the model’s guardrails are not your control, and the controls that still hold when an agent moves in minutes.

Fact check: 10 claims checked (5 verified, 2 partly right, 1 disputed, 2 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/crowdstrike-uncovers-new-agentic-ai-campaigns/</description>
<enclosure url="https://podcast.agile-insights.com.au/crowdstrike-uncovers-new-agentic-ai-campaigns/primer-natural.mp3" length="5045393" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/crowdstrike-uncovers-new-agentic-ai-campaigns/cover.jpg?v=7c2f8d0a"/>
<category>security</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 28: When an AI keeps the wiki, checking has to keep pace with writing · Part 2: What compounds, and who checks</title>
<link>https://podcast.agile-insights.com.au/2611-llm-wiki-checking-keeps-pace/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2611-llm-wiki-checking-keeps-pace/discussion.mp3</guid>
<pubDate>Fri, 06 Nov 2026 06:28:01 +1100</pubDate>
<description>Andrej Karpathy’s L L M wiki has an agent compile your sources into linked pages and keep them current. What the pattern is, why errors compound as well as insight, and four controls for a knowledge base your agents write for themselves.

Fact check: 10 claims checked (5 confirmed, 2 reported, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2611-llm-wiki-checking-keeps-pace/</description>
<enclosure url="https://podcast.agile-insights.com.au/2611-llm-wiki-checking-keeps-pace/discussion.mp3" length="6051016" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2611-llm-wiki-checking-keeps-pace/cover.jpg?v=d6a5891b"/>
<category>data</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 28: When an AI keeps the wiki, checking has to keep pace with writing · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/2611-llm-wiki-checking-keeps-pace/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2611-llm-wiki-checking-keeps-pace/primer.mp3</guid>
<pubDate>Fri, 06 Nov 2026 06:28:00 +1100</pubDate>
<description>Andrej Karpathy’s L L M wiki has an agent compile your sources into linked pages and keep them current. What the pattern is, why errors compound as well as insight, and four controls for a knowledge base your agents write for themselves.

Fact check: 10 claims checked (5 confirmed, 2 reported, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2611-llm-wiki-checking-keeps-pace/</description>
<enclosure url="https://podcast.agile-insights.com.au/2611-llm-wiki-checking-keeps-pace/primer.mp3" length="5075031" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2611-llm-wiki-checking-keeps-pace/cover.jpg?v=d6a5891b"/>
<category>data</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 27: Can you trust an AI judge? · Part 2: Discussion</title>
<link>https://podcast.agile-insights.com.au/2609-ai-judgment-models/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2609-ai-judgment-models/discussion-natural.mp3</guid>
<pubDate>Thu, 05 Nov 2026 06:27:01 +1100</pubDate>
<description>An AI judge is a configuration, not an instrument. Ai2 found the same model gives opposite verdicts on the same cases when the prompt changes, so test the whole set-up on the cases it will actually judge before trusting its score.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2609-ai-judgment-models/</description>
<enclosure url="https://podcast.agile-insights.com.au/2609-ai-judgment-models/discussion-natural.mp3" length="5828407" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2609-ai-judgment-models/cover.jpg?v=cdbdac3c"/>
<category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 27: Can you trust an AI judge? · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/2609-ai-judgment-models/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2609-ai-judgment-models/primer-natural.mp3</guid>
<pubDate>Thu, 05 Nov 2026 06:27:00 +1100</pubDate>
<description>An AI judge is a configuration, not an instrument. Ai2 found the same model gives opposite verdicts on the same cases when the prompt changes, so test the whole set-up on the cases it will actually judge before trusting its score.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2609-ai-judgment-models/</description>
<enclosure url="https://podcast.agile-insights.com.au/2609-ai-judgment-models/primer-natural.mp3" length="4922909" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2609-ai-judgment-models/cover.jpg?v=cdbdac3c"/>
<category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 26: The agent that speaks first: what to decide before proactive agents reach work accounts · Part 2: Discussion</title>
<link>https://podcast.agile-insights.com.au/ai-native-organisation/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ai-native-organisation/discussion-natural.mp3</guid>
<pubDate>Tue, 03 Nov 2026 06:26:01 +1100</pubDate>
<description>Anthropic&#x27;s Claude Duet messages you first, checks your accounts and asks before it acts. An agent that comes to you will chase what it can reach, not what matters most, so governing one is about priorities, approvals and an off switch.

Fact check: 10 claims checked (7 confirmed, 2 reported, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ai-native-organisation/</description>
<enclosure url="https://podcast.agile-insights.com.au/ai-native-organisation/discussion-natural.mp3" length="5226424" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ai-native-organisation/cover.jpg?v=23f50c68"/>
<category>agents</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 26: The agent that speaks first: what to decide before proactive agents reach work accounts · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ai-native-organisation/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ai-native-organisation/primer-natural.mp3</guid>
<pubDate>Tue, 03 Nov 2026 06:26:00 +1100</pubDate>
<description>Anthropic&#x27;s Claude Duet messages you first, checks your accounts and asks before it acts. An agent that comes to you will chase what it can reach, not what matters most, so governing one is about priorities, approvals and an off switch.

Fact check: 10 claims checked (7 confirmed, 2 reported, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ai-native-organisation/</description>
<enclosure url="https://podcast.agile-insights.com.au/ai-native-organisation/primer-natural.mp3" length="4702542" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ai-native-organisation/cover.jpg?v=23f50c68"/>
<category>agents</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 25: The managed loop: what you still own when the vendor runs your agent · Part 2: Discussion</title>
<link>https://podcast.agile-insights.com.au/2610-foundry-copilot-harness-managed-agent-loop/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-foundry-copilot-harness-managed-agent-loop/discussion-natural.mp3</guid>
<pubDate>Mon, 02 Nov 2026 06:25:01 +1100</pubDate>
<description>Microsoft now runs the agent loop and sandbox for you in Foundry. You hand over the plumbing but not the accountability: the skills, the tool settings and what happens on a restart stay yours, and in the preview they start wide open.

Fact check: 12 claims checked (9 verified, 1 partly right, 1 disputed, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-foundry-copilot-harness-managed-agent-loop/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-foundry-copilot-harness-managed-agent-loop/discussion-natural.mp3" length="5750930" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-foundry-copilot-harness-managed-agent-loop/cover.jpg?v=47968fb9"/>
<category>microsoft</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 25: The managed loop: what you still own when the vendor runs your agent · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/2610-foundry-copilot-harness-managed-agent-loop/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-foundry-copilot-harness-managed-agent-loop/primer-natural.mp3</guid>
<pubDate>Mon, 02 Nov 2026 06:25:00 +1100</pubDate>
<description>Microsoft now runs the agent loop and sandbox for you in Foundry. You hand over the plumbing but not the accountability: the skills, the tool settings and what happens on a restart stay yours, and in the preview they start wide open.

Fact check: 12 claims checked (9 verified, 1 partly right, 1 disputed, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-foundry-copilot-harness-managed-agent-loop/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-foundry-copilot-harness-managed-agent-loop/primer-natural.mp3" length="4655481" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-foundry-copilot-harness-managed-agent-loop/cover.jpg?v=47968fb9"/>
<category>microsoft</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 24: Remember, retrieve, recheck, forget: governing agent memory</title>
<link>https://podcast.agile-insights.com.au/2610-agent-memory-remember-retrieve-revalidate-forget/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-agent-memory-remember-retrieve-revalidate-forget/primer-natural.mp3</guid>
<pubDate>Mon, 02 Nov 2026 06:24:00 +1100</pubDate>
<description>Agent memory is not one feature but four decisions: what to remember, who may retrieve it, when to recheck it against the real record, and how to make it truly forget. None of them should be left to the agent.

Fact check: 12 claims checked (7 verified, 4 partly right, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-agent-memory-remember-retrieve-revalidate-forget/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-agent-memory-remember-retrieve-revalidate-forget/primer-natural.mp3" length="5258980" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-agent-memory-remember-retrieve-revalidate-forget/cover.jpg?v=6ad04c1a"/>
<category>agents</category><category>data</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 23: What FDE means at OpenAI: skills, not custom code · Part 2: Discussion</title>
<link>https://podcast.agile-insights.com.au/2610-forward-deployed-engineering-skills-not-custom-code/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-forward-deployed-engineering-skills-not-custom-code/discussion.mp3</guid>
<pubDate>Wed, 28 Oct 2026 06:23:01 +1100</pubDate>
<description>When an AI lab&#x27;s engineers build your AI well, what you keep is a shared library of skills and the evals that prove them. Four terms, the case for taking the help, and what to own before the engineers arrive.

Fact check: 12 claims checked (3 confirmed, 4 reported, 2 partly right, 1 disputed, 2 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-forward-deployed-engineering-skills-not-custom-code/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-forward-deployed-engineering-skills-not-custom-code/discussion.mp3" length="4984482" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-forward-deployed-engineering-skills-not-custom-code/cover.jpg?v=54da9b59"/>
<category>agents</category><category>leadership</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 23: What FDE means at OpenAI: skills, not custom code · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/2610-forward-deployed-engineering-skills-not-custom-code/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-forward-deployed-engineering-skills-not-custom-code/primer.mp3</guid>
<pubDate>Wed, 28 Oct 2026 06:23:00 +1100</pubDate>
<description>When an AI lab&#x27;s engineers build your AI well, what you keep is a shared library of skills and the evals that prove them. Four terms, the case for taking the help, and what to own before the engineers arrive.

Fact check: 12 claims checked (3 confirmed, 4 reported, 2 partly right, 1 disputed, 2 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-forward-deployed-engineering-skills-not-custom-code/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-forward-deployed-engineering-skills-not-custom-code/primer.mp3" length="4966319" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-forward-deployed-engineering-skills-not-custom-code/cover.jpg?v=54da9b59"/>
<category>agents</category><category>leadership</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 22: From Pilots to Programs: govern an agent by what it does · Part 2: Discussion</title>
<link>https://podcast.agile-insights.com.au/2610-microsoft-pilots-to-programs-agentic-coe/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-microsoft-pilots-to-programs-agentic-coe/discussion.mp3</guid>
<pubDate>Wed, 28 Oct 2026 06:22:01 +1100</pubDate>
<description>Microsoft argues that AI pilots stall on the operating model, not the technology. The line between assisting and acting, three tiers of oversight, a two week test for friction, and how much of a vendor’s framework to take.

Fact check: 9 claims checked (4 confirmed, 1 reported, 2 partly right, 2 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-microsoft-pilots-to-programs-agentic-coe/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-microsoft-pilots-to-programs-agentic-coe/discussion.mp3" length="5692849" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-microsoft-pilots-to-programs-agentic-coe/cover.jpg?v=beb90ac0"/>
<category>microsoft</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 22: From Pilots to Programs: govern an agent by what it does · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/2610-microsoft-pilots-to-programs-agentic-coe/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-microsoft-pilots-to-programs-agentic-coe/primer.mp3</guid>
<pubDate>Wed, 28 Oct 2026 06:22:00 +1100</pubDate>
<description>Microsoft argues that AI pilots stall on the operating model, not the technology. The line between assisting and acting, three tiers of oversight, a two week test for friction, and how much of a vendor’s framework to take.

Fact check: 9 claims checked (4 confirmed, 1 reported, 2 partly right, 2 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-microsoft-pilots-to-programs-agentic-coe/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-microsoft-pilots-to-programs-agentic-coe/primer.mp3" length="5032836" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-microsoft-pilots-to-programs-agentic-coe/cover.jpg?v=beb90ac0"/>
<category>microsoft</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 21: Trust cannot be rushed: how much authority will you give AI?</title>
<link>https://podcast.agile-insights.com.au/trust-cannot-be-rushed/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/trust-cannot-be-rushed/audio.mp3</guid>
<pubDate>Tue, 27 Oct 2026 06:21:00 +1100</pubDate>
<description>Mike Morgan&#x27;s argument that trust is the authority you grant AI, what the Australian law already requires from 10 December 2026, and how much of the speed claim the evidence carries.

Fact check: 6 claims checked (3 confirmed, 2 partly right, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/trust-cannot-be-rushed/</description>
<enclosure url="https://podcast.agile-insights.com.au/trust-cannot-be-rushed/audio.mp3" length="7204562" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/trust-cannot-be-rushed/cover.jpg?v=12c56e5b"/>
<category>leadership</category><category>rules</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 20: Usage is not quality: who reads the feedback? · Part 2: Who should see the feedback?</title>
<link>https://podcast.agile-insights.com.au/explore-user-feedback-with-microsoft-365-insights-agent/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/explore-user-feedback-with-microsoft-365-insights-agent/discussion.mp3</guid>
<pubDate>Tue, 27 Oct 2026 06:20:01 +1100</pubDate>
<description>Microsoft&#x27;s Insights Agent feedback preview shows what users say about your agents. Why usage counts are not a quality measure, why the AI theme summary is not the evidence, and what to decide before you switch it on.

Fact check: 18 claims checked (14 verified, 1 partly right, 3 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/explore-user-feedback-with-microsoft-365-insights-agent/</description>
<enclosure url="https://podcast.agile-insights.com.au/explore-user-feedback-with-microsoft-365-insights-agent/discussion.mp3" length="5627333" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/explore-user-feedback-with-microsoft-365-insights-agent/cover.jpg?v=440c93a8"/>
<category>microsoft</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 20: Usage is not quality: who reads the feedback? · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/explore-user-feedback-with-microsoft-365-insights-agent/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/explore-user-feedback-with-microsoft-365-insights-agent/primer.mp3</guid>
<pubDate>Tue, 27 Oct 2026 06:20:00 +1100</pubDate>
<description>Microsoft&#x27;s Insights Agent feedback preview shows what users say about your agents. Why usage counts are not a quality measure, why the AI theme summary is not the evidence, and what to decide before you switch it on.

Fact check: 18 claims checked (14 verified, 1 partly right, 3 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/explore-user-feedback-with-microsoft-365-insights-agent/</description>
<enclosure url="https://podcast.agile-insights.com.au/explore-user-feedback-with-microsoft-365-insights-agent/primer.mp3" length="4577322" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/explore-user-feedback-with-microsoft-365-insights-agent/cover.jpg?v=440c93a8"/>
<category>microsoft</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 19: A read-only instruction is a request, not a control · Part 2: Whose failure was it?</title>
<link>https://podcast.agile-insights.com.au/2610-openai-model-breached-dashboard-owner-systems/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-openai-model-breached-dashboard-owner-systems/discussion.mp3</guid>
<pubDate>Mon, 26 Oct 2026 06:19:01 +1100</pubDate>
<description>OpenAI reports that a model told to only look at a public dashboard changed a live service’s settings and kept using access it gained. Why instructions are not controls, why undoing is not cancelling, and why the agent’s own account is not the record.

Fact check: 8 claims checked (3 confirmed, 2 reported, 2 partly right, 1 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-openai-model-breached-dashboard-owner-systems/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-openai-model-breached-dashboard-owner-systems/discussion.mp3" length="5570222" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-openai-model-breached-dashboard-owner-systems/cover.jpg?v=cd2e122d"/>
<category>agents</category><category>security</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 19: A read-only instruction is a request, not a control · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/2610-openai-model-breached-dashboard-owner-systems/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/2610-openai-model-breached-dashboard-owner-systems/primer.mp3</guid>
<pubDate>Mon, 26 Oct 2026 06:19:00 +1100</pubDate>
<description>OpenAI reports that a model told to only look at a public dashboard changed a live service’s settings and kept using access it gained. Why instructions are not controls, why undoing is not cancelling, and why the agent’s own account is not the record.

Fact check: 8 claims checked (3 confirmed, 2 reported, 2 partly right, 1 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/2610-openai-model-breached-dashboard-owner-systems/</description>
<enclosure url="https://podcast.agile-insights.com.au/2610-openai-model-breached-dashboard-owner-systems/primer.mp3" length="4841553" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/2610-openai-model-breached-dashboard-owner-systems/cover.jpg?v=cd2e122d"/>
<category>agents</category><category>security</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 18: When agents share one memory, memory is a data flow to govern · Part 2: Convenience, suppliers and what we do not know</title>
<link>https://podcast.agile-insights.com.au/ep18/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep18/discussion.mp3</guid>
<pubDate>Mon, 26 Oct 2026 06:18:01 +1100</pubDate>
<description>An open source tool called Claude Mem lets coding agents from different companies read and write the same memory. What agent memory is, why local is not the same as private, and the questions to answer before anyone connects a second agent.

Fact check: 16 claims checked (13 verified, 1 partly right, 2 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep18/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep18/discussion.mp3" length="5636185" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep18/cover.jpg?v=e40b23f1"/>
<category>agents</category><category>data</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 18: When agents share one memory, memory is a data flow to govern · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ep18/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep18/primer.mp3</guid>
<pubDate>Mon, 26 Oct 2026 06:18:00 +1100</pubDate>
<description>An open source tool called Claude Mem lets coding agents from different companies read and write the same memory. What agent memory is, why local is not the same as private, and the questions to answer before anyone connects a second agent.

Fact check: 16 claims checked (13 verified, 1 partly right, 2 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep18/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep18/primer.mp3" length="4741430" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep18/cover.jpg?v=e40b23f1"/>
<category>agents</category><category>data</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 17: Switching on a role pack is a process decision, not an I T toggle · Part 2: Turn them on, or slow them down?</title>
<link>https://podcast.agile-insights.com.au/ep17/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep17/discussion.mp3</guid>
<pubDate>Mon, 26 Oct 2026 06:17:01 +1100</pubDate>
<description>Microsoft has added ready made Sales, Service, Education and Legal skills to Copilot Cowork. What a skill is, why your own playbook becomes the control, where the admin switch stops, and what to decide before people start using them.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep17/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep17/discussion.mp3" length="5349687" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep17/cover.jpg?v=e0aba3d9"/>
<category>microsoft</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 17: Switching on a role pack is a process decision, not an I T toggle · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ep17/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep17/primer.mp3</guid>
<pubDate>Mon, 26 Oct 2026 06:17:00 +1100</pubDate>
<description>Microsoft has added ready made Sales, Service, Education and Legal skills to Copilot Cowork. What a skill is, why your own playbook becomes the control, where the admin switch stops, and what to decide before people start using them.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep17/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep17/primer.mp3" length="4865083" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep17/cover.jpg?v=e0aba3d9"/>
<category>microsoft</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 16: When making something gets cheap, checking it becomes the real work · Part 2: Should you encourage orthogonal work?</title>
<link>https://podcast.agile-insights.com.au/ep16/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep16/discussion.mp3</guid>
<pubDate>Fri, 23 Oct 2026 06:16:01 +1100</pubDate>
<description>A I is changing what work an organisation can afford to take on, not only how fast it does its usual work. What an investment firm’s scorpion toxin experiment shows about orthogonal work, the cost of checking, and where the outputs go.

Fact check: 11 claims checked (4 confirmed, 5 reported, 1 disputed), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep16/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep16/discussion.mp3" length="5115177" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep16/cover.jpg?v=c81b9a7b"/>
<category>oversight</category><category>leadership</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 16: When making something gets cheap, checking it becomes the real work · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ep16/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep16/primer.mp3</guid>
<pubDate>Fri, 23 Oct 2026 06:16:00 +1100</pubDate>
<description>A I is changing what work an organisation can afford to take on, not only how fast it does its usual work. What an investment firm’s scorpion toxin experiment shows about orthogonal work, the cost of checking, and where the outputs go.

Fact check: 11 claims checked (4 confirmed, 5 reported, 1 disputed), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep16/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep16/primer.mp3" length="4428041" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep16/cover.jpg?v=c81b9a7b"/>
<category>oversight</category><category>leadership</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 15: Microsoft governs the access. You still have to govern the answer · Part 2: Is inherited governance enough?</title>
<link>https://podcast.agile-insights.com.au/ep15/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep15/discussion.mp3</guid>
<pubDate>Fri, 23 Oct 2026 06:15:01 +1100</pubDate>
<description>Copilot Code lets anyone in the Frontier preview build an app with A I and share it with colleagues. What Microsoft’s managed runtime governs, what it leaves to the organisation, and the defaults to decide now.

Fact check: 26 claims checked (17 verified, 3 partly right, 1 disputed, 3 unverified, 2 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep15/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep15/discussion.mp3" length="5620021" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep15/cover.jpg?v=13940f58"/>
<category>microsoft</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 15: Microsoft governs the access. You still have to govern the answer · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ep15/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep15/primer.mp3</guid>
<pubDate>Fri, 23 Oct 2026 06:15:00 +1100</pubDate>
<description>Copilot Code lets anyone in the Frontier preview build an app with A I and share it with colleagues. What Microsoft’s managed runtime governs, what it leaves to the organisation, and the defaults to decide now.

Fact check: 26 claims checked (17 verified, 3 partly right, 1 disputed, 3 unverified, 2 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep15/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep15/primer.mp3" length="4832735" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep15/cover.jpg?v=13940f58"/>
<category>microsoft</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 14: Write down what the agent must never do, and what it must still do · Part 2: The contract, then the test</title>
<link>https://podcast.agile-insights.com.au/ep14/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep14/discussion.mp3</guid>
<pubDate>Wed, 21 Oct 2026 06:14:01 +1100</pubDate>
<description>Why every AI agent needs a short behaviour contract before you build or buy it, what that one page does for governance, and how Microsoft’s open source ASSERT tests an agent against it.

Fact check: 18 claims checked (14 confirmed, 3 reported, 1 partly right), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep14/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep14/discussion.mp3" length="7135291" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep14/cover.jpg?v=4c3c8e7f"/>
<category>agents</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 14: Write down what the agent must never do, and what it must still do · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ep14/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep14/primer.mp3</guid>
<pubDate>Wed, 21 Oct 2026 06:14:00 +1100</pubDate>
<description>Why every AI agent needs a short behaviour contract before you build or buy it, what that one page does for governance, and how Microsoft’s open source ASSERT tests an agent against it.

Fact check: 18 claims checked (14 confirmed, 3 reported, 1 partly right), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep14/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep14/primer.mp3" length="4674070" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep14/cover.jpg?v=4c3c8e7f"/>
<category>agents</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 13: Fairness under pressure: a fair AI decision has to hold when someone leans on it · Part 2: What a pressure test changes</title>
<link>https://podcast.agile-insights.com.au/ep13/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep13/discussion.mp3</guid>
<pubDate>Wed, 21 Oct 2026 06:13:01 +1100</pubDate>
<description>New research from Ai2 shows AI models that pass the usual name-swap test can still give way unevenly under one sentence of pressure. What that means for any organisation using AI to choose between people.

Fact check: 18 claims checked (17 confirmed, 1 partly right), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep13/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep13/discussion.mp3" length="5697517" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep13/cover.jpg?v=ad303f7e"/>
<category>people</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 13: Fairness under pressure: a fair AI decision has to hold when someone leans on it · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ep13/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep13/primer.mp3</guid>
<pubDate>Wed, 21 Oct 2026 06:13:00 +1100</pubDate>
<description>New research from Ai2 shows AI models that pass the usual name-swap test can still give way unevenly under one sentence of pressure. What that means for any organisation using AI to choose between people.

Fact check: 18 claims checked (17 confirmed, 1 partly right), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep13/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep13/primer.mp3" length="4527998" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep13/cover.jpg?v=ad303f7e"/>
<category>people</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 12: Who reads the code when the AI writes it? · Part 2: Is it safe to stop reading?</title>
<link>https://podcast.agile-insights.com.au/ep12/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep12/discussion.mp3</guid>
<pubDate>Wed, 21 Oct 2026 06:12:01 +1100</pubDate>
<description>Why asking whether a person looked at it is a weak control for AI written work, and how to match the check to the risk, keep people on the risky calls and make mistakes cheap to undo.

Fact check: 25 claims checked (15 confirmed, 9 reported, 1 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep12/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep12/discussion.mp3" length="5898129" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep12/cover.jpg?v=885e5eb9"/>
<category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 12: Who reads the code when the AI writes it? · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ep12/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep12/primer.mp3</guid>
<pubDate>Wed, 21 Oct 2026 06:12:00 +1100</pubDate>
<description>Why asking whether a person looked at it is a weak control for AI written work, and how to match the check to the risk, keep people on the risky calls and make mistakes cheap to undo.

Fact check: 25 claims checked (15 confirmed, 9 reported, 1 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep12/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep12/primer.mp3" length="4922928" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep12/cover.jpg?v=885e5eb9"/>
<category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 11: Agents are insiders. The controls that matter sit outside them · Part 2: Where the controls sit</title>
<link>https://podcast.agile-insights.com.au/ep11/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep11/discussion.mp3</guid>
<pubDate>Wed, 21 Oct 2026 06:11:01 +1100</pubDate>
<description>Why AI agents go wrong without any malice, why a list of rules is not a boundary, and why safe autonomy depends on where the controls sit, not how many you add.

Fact check: 21 claims checked (10 confirmed, 8 reported, 1 disputed, 2 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep11/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep11/discussion.mp3" length="6024847" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep11/cover.jpg?v=b8e77759"/>
<category>agents</category><category>security</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 11: Agents are insiders. The controls that matter sit outside them · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ep11/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep11/primer.mp3</guid>
<pubDate>Wed, 21 Oct 2026 06:11:00 +1100</pubDate>
<description>Why AI agents go wrong without any malice, why a list of rules is not a boundary, and why safe autonomy depends on where the controls sit, not how many you add.

Fact check: 21 claims checked (10 confirmed, 8 reported, 1 disputed, 2 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep11/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep11/primer.mp3" length="4935910" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep11/cover.jpg?v=b8e77759"/>
<category>agents</category><category>security</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 10: Microsoft now ships the controls for AI agents. The decisions are still yours · Part 2: What each one fixes</title>
<link>https://podcast.agile-insights.com.au/ep10/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep10/ep10-part2-discussion.mp3</guid>
<pubDate>Fri, 16 Oct 2026 06:10:01 +1100</pubDate>
<description>What Microsoft Build 2026 turned into products, the agent problem each one fixes, and what your organisation still has to decide for the fix to hold.

Fact check: 16 claims checked (15 confirmed, 1 reported), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep10/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep10/ep10-part2-discussion.mp3" length="6275586" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep10/cover.jpg?v=e8f4ce47"/>
<category>microsoft</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 10: Microsoft now ships the controls for AI agents. The decisions are still yours · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ep10/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep10/ep10-part1-primer.mp3</guid>
<pubDate>Fri, 16 Oct 2026 06:10:00 +1100</pubDate>
<description>What Microsoft Build 2026 turned into products, the agent problem each one fixes, and what your organisation still has to decide for the fix to hold.

Fact check: 16 claims checked (15 confirmed, 1 reported), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep10/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep10/ep10-part1-primer.mp3" length="4362352" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep10/cover.jpg?v=e8f4ce47"/>
<category>microsoft</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 9: Your company brain is a copy. Govern it like one</title>
<link>https://podcast.agile-insights.com.au/ep9/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep9/script.mp3</guid>
<pubDate>Mon, 12 Oct 2026 06:09:00 +1100</pubDate>
<description>A company brain is a copy of your data and a copy of your rules. Before you build one, decide who owns each copy, who may see it, how long it lives and who checks what it says.

Fact check: 18 claims checked (11 confirmed, 4 reported, 1 partly right, 1 unverified, 1 not yet known), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep9/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep9/script.mp3" length="5459837" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep9/cover.jpg?v=5bf553bb"/>
<category>data</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 8: The Agreeable Machine: when staff think problems through with AI</title>
<link>https://podcast.agile-insights.com.au/ep8/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep8/script.mp3</guid>
<pubDate>Sun, 11 Oct 2026 06:08:00 +1100</pubDate>
<description>When people talk their own problems through with AI, the main risk is not bad information. It is that the machine agrees with them.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep8/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep8/script.mp3" length="7316639" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep8/cover.jpg?v=362b895e"/>
<category>people</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 7: The Tool That Never Says Stop: AI, ADHD and what an organisation owes its staff</title>
<link>https://podcast.agile-insights.com.au/ep7/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep7/script.mp3</guid>
<pubDate>Sun, 11 Oct 2026 06:07:00 +1100</pubDate>
<description>AI helps people with ADHD most with starting work and getting back into it. The same design takes away every cue to stop, and the evidence is thin on both counts.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep7/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep7/script.mp3" length="7132849" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep7/cover.jpg?v=aa2ef27d"/>
<category>people</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 6: The Five Dysfunctions of a Team: a checklist for AI adoption, not a law</title>
<link>https://podcast.agile-insights.com.au/ep6/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep6/script.mp3</guid>
<pubDate>Sat, 10 Oct 2026 06:06:00 +1100</pubDate>
<description>Lencioni&#x27;s pyramid is a useful prompt for the human side of AI programmes. The trust layer has research behind it. The strict order does not.

Fact check: 8 claims checked (3 confirmed, 3 reported, 1 disputed, 1 unverified), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep6/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep6/script.mp3" length="6520369" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep6/cover.jpg?v=31bef4ec"/>
<category>leadership</category><category>people</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 5: Are data centres evil?</title>
<link>https://podcast.agile-insights.com.au/ep5/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep5/script.mp3</guid>
<pubDate>Sat, 10 Oct 2026 06:05:00 +1100</pubDate>
<description>AI&#x27;s power, water and carbon footprint is a trade-off an executive can manage, not a verdict.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep5/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep5/script.mp3" length="3829880" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep5/cover.jpg?v=35e07d48"/>
<category>leadership</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 4: Is your digital workplace a safety risk? AI, monitoring and work health and safety</title>
<link>https://podcast.agile-insights.com.au/ep4/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep4/script.mp3</guid>
<pubDate>Sat, 10 Oct 2026 06:04:00 +1100</pubDate>
<description>In New South Wales, software that allocates, measures and monitors work is now a safety matter, and most of the new duties start soon.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep4/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep4/script.mp3" length="3078250" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep4/cover.jpg?v=8914b35f"/>
<category>rules</category><category>people</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 3: Is your AI making decisions about people? What the Privacy Act now asks of you</title>
<link>https://podcast.agile-insights.com.au/ep3/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep3/script.mp3</guid>
<pubDate>Sat, 10 Oct 2026 06:03:00 +1100</pubDate>
<description>From 10 December 2026, organisations must say in their privacy policy when software makes or shapes significant decisions about people.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep3/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep3/script.mp3" length="3543672" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep3/cover.jpg?v=3ef9b552"/>
<category>rules</category><category>people</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 2: AI escaping containment? What to check this week · Part 2: Discussion</title>
<link>https://podcast.agile-insights.com.au/ep2/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep2/ep2-part2-discussion.mp3</guid>
<pubDate>Sat, 10 Oct 2026 06:02:01 +1100</pubDate>
<description>What the reports about AI agents and boundaries actually show, how to read the Medicare portal incident, and the checks an organisation can make now.

Fact check: 15 claims checked (3 confirmed, 9 reported, 1 disputed), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep2/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep2/ep2-part2-discussion.mp3" length="4185409" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep2/cover.jpg?v=8b173eec"/>
<category>security</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 2: AI escaping containment? What to check this week · Part 1: Primer</title>
<link>https://podcast.agile-insights.com.au/ep2/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep2/ep2-part1-primer.mp3</guid>
<pubDate>Sat, 10 Oct 2026 06:02:00 +1100</pubDate>
<description>What the reports about AI agents and boundaries actually show, how to read the Medicare portal incident, and the checks an organisation can make now.

Fact check: 15 claims checked (3 confirmed, 9 reported, 1 disputed), in the show notes.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep2/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep2/ep2-part1-primer.mp3" length="3685727" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep2/cover.jpg?v=8b173eec"/>
<category>security</category><category>agents</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item><item>
<title>Episode 1: Staying in charge of decisions in the age of AI</title>
<link>https://podcast.agile-insights.com.au/ep1/</link>
<guid isPermaLink="false">https://podcast.agile-insights.com.au/ep1/audio.mp3</guid>
<pubDate>Sat, 10 Oct 2026 06:01:00 +1100</pubDate>
<description>Dr Chris Stephenson&#x27;s Decision Equilibrium, and what an organisation needs when AI starts making decisions.

Show notes, sources and transcript: https://podcast.agile-insights.com.au/ep1/</description>
<enclosure url="https://podcast.agile-insights.com.au/ep1/audio.mp3" length="5309479" type="audio/mpeg"/>

<itunes:image href="https://podcast.agile-insights.com.au/ep1/cover.jpg?v=94309214"/>
<category>leadership</category><category>oversight</category>
<itunes:episodeType>full</itunes:episodeType>
<itunes:explicit>false</itunes:explicit>
</item>
</channel>
</rss>
