<![CDATA[Agent Cribs]]>Agent Cribs takes you inside the AI agents people are actually using at work. Each episode breaks down a real workflow: what the agent does, how it was built, where human judgment matters, and what you can copy for your own team. No vague predictions, just practical systems, honest lessons, and agents doing real work.

Hosted by Ricardo Ghekiere, Founder of Patricia.app the AI Coworker for Small Teams.

]]>https://www.patricia.app/Riverside.fm (https://riverside.com)Fri, 11 Sep 2026 05:13:41 GMTWed, 19 Aug 2026 08:43:16 GMT60Ricardo Ghekiere<p>Agent Cribs takes you inside the AI agents people are actually using at work. Each episode breaks down a real workflow: what the agent does, how it was built, where human judgment matters, and what you can copy for your own team. No vague predictions, just practical systems, honest lessons, and agents doing real work. </p><p></p><p>Hosted by Ricardo Ghekiere, Founder of <a rel="noopener noreferrer nofollow" href="http://Patricia.app" target="_blank">Patricia.app</a> the AI Coworker for Small Teams.</p>episodicRicardo Ghekierericardo@betterpic.iono<![CDATA[His 11 AI Agents Made Him 20 to 25x More Productive with Bhanu Chawla]]>In this episode of Agent Cribs he opens Frame Game, the retro control room he built for his agents. They design ads, write copy, launch Meta campaigns, debate ideas in a boardroom, and report back to him for approval. All of it runs on a $200 Claude subscription on a Mac Mini.

Chapters:
0:00 Introduction to Bhanu and his AI system
0:30 Who is Bhanu Chawla
1:49 Overview of the AI agent setup and purpose
2:35 Frame Game demo: welcome to the crib
3:30 Demonstration of the AI agents' capabilities
6:27 Design choices: multiple agents vs single agent
8:16 How agents communicate and collaborate
10:48 Triaging and automating marketing campaigns
13:41 Running campaigns autonomously and cost management
16:24 Feedback, reworking, and learning in agents
18:09 Time spent reviewing and productivity gains
20:28 Running multi-agent debates and brainstorming
23:35 Hiring, onboarding, and managing AI agents
26:43 AI brings speed, humans bring taste
28:13 Onboarding a human into the agent system
31:18 Build the interface that works for you
33:09 Future plans and lessons learned
34:37 How long to train an AI employee
36:03 Wrap up Agent Cribs is the show where operators open the door and show you the agents they actually run, instead of the ones they talk about.

Hosted by Ricardo Ghekiere, founder of Patricia, the AI teammate for Slack and Microsoft Teams.

]]>1e49019c-99c3-4c10-8f61-acf72ee46ad0Thu, 03 Sep 2026 13:05:44 GMT<p>In this episode of Agent Cribs he opens Frame Game, the retro control room he built for his agents. They design ads, write copy, launch Meta campaigns, debate ideas in a boardroom, and report back to him for approval. All of it runs on a $200 Claude subscription on a Mac Mini. <br /><br />Chapters: <br />0:00 Introduction to Bhanu and his AI system <br />0:30 Who is Bhanu Chawla <br />1:49 Overview of the AI agent setup and purpose <br />2:35 Frame Game demo: welcome to the crib <br />3:30 Demonstration of the AI agents' capabilities <br />6:27 Design choices: multiple agents vs single agent <br />8:16 How agents communicate and collaborate <br />10:48 Triaging and automating marketing campaigns <br />13:41 Running campaigns autonomously and cost management <br />16:24 Feedback, reworking, and learning in agents <br />18:09 Time spent reviewing and productivity gains <br />20:28 Running multi-agent debates and brainstorming <br />23:35 Hiring, onboarding, and managing AI agents <br />26:43 AI brings speed, humans bring taste <br />28:13 Onboarding a human into the agent system <br />31:18 Build the interface that works for you <br />33:09 Future plans and lessons learned <br />34:37 How long to train an AI employee <br />36:03 Wrap up Agent Cribs is the show where operators open the door and show you the agents they actually run, instead of the ones they talk about. <br /><br />Hosted by Ricardo Ghekiere, founder of Patricia, the AI teammate for Slack and Microsoft Teams.</p>no00:36:522His 11 AI Agents Made Him 20 to 25x More Productive with Bhanu Chawlafull<![CDATA[Building an AI Agent for High-Quality Blog Posts with Curran Dye]]>Curran Dye leads marketing at Supy, the AI inventory platform used by multi site restaurant groups. In this first episode of Agent Cribs he shares his screen and shows the SEO agent that has written about 450 blog posts, work that used to cost roughly $75 a post from outside writers. That is about $30,000 of writing budget, and by his own count closer to $40,000 to $50,000 once you price the clicks.

No slides. Real screens:

  • The GitHub repo he treats as the agent's brain, one repo per agent, with the skills, the briefs and a process diagram in the README.
  • Two internal MCP servers. Nexus holds every sales call and customer conversation, so the agent knows what buyers actually ask. Cortex reads the Supy codebase, so the agent never claims a feature the product does not have.
  • Claude routines doing the work while his laptop is shut, and Supabase holding the drafts.
  • A Slack channel per agent, where the post scores itself out of 10 against 20 checks, then waits for approve or request edits before it goes live on Webflow.

Ricardo also pushes on the awkward parts. What happens to the agent if Curran gets hit by a bus. Whether the repo belongs to him or to Supy. Whether a marketer should even want the whole team to run it. And what broke on the way here, including the weeks when the laptop had to stay open in his bag for the job to run.

Curran closes with two pieces of advice for anyone building their first real agent: be inquisitive rather than lazy, and read your own files line by line, because the output is only ever as good as the input.

]]>ff3d4f00-9cc0-4c70-8f2f-656fada7b0aeWed, 19 Aug 2026 14:19:58 GMT<p>Curran Dye leads marketing at Supy, the AI inventory platform used by multi site restaurant groups. In this first episode of Agent Cribs he shares his screen and shows the SEO agent that has written about 450 blog posts, work that used to cost roughly $75 a post from outside writers. That is about $30,000 of writing budget, and by his own count closer to $40,000 to $50,000 once you price the clicks.</p><p></p><p>No slides. Real screens:</p><blockquote><ul><li>The GitHub repo he treats as the agent's brain, one repo per agent, with the skills, the briefs and a process diagram in the README.</li><li>Two internal MCP servers. Nexus holds every sales call and customer conversation, so the agent knows what buyers actually ask. Cortex reads the Supy codebase, so the agent never claims a feature the product does not have.</li><li>Claude routines doing the work while his laptop is shut, and Supabase holding the drafts.</li><li>A Slack channel per agent, where the post scores itself out of 10 against 20 checks, then waits for approve or request edits before it goes live on Webflow.</li></ul><p></p><p>Ricardo also pushes on the awkward parts. What happens to the agent if Curran gets hit by a bus. Whether the repo belongs to him or to Supy. Whether a marketer should even want the whole team to run it. And what broke on the way here, including the weeks when the laptop had to stay open in his bag for the job to run.</p><p></p><p>Curran closes with two pieces of advice for anyone building their first real agent: be inquisitive rather than lazy, and read your own files line by line, because the output is only ever as good as the input.</p><p></p></blockquote>no00:26:4611Building an AI Agent for High-Quality Blog Posts with Curran Dyefull