Episode 438: Taking your WorkOS and LLM Wiki to the Next Level

Welcome to Episode 438 of the Microsoft Cloud IT Pro Podcast. Back in Episode 429, Ben and Scott introduced LLM wikis. In this episode, they come back a few months later to compare what those wikis turned into after real daily use: personalized work operating systems, or a WorkOS, that tie together tasks, meetings, research, communications, and recurring workflows. Scott walks through the skills that now run his day, from a 9:00 a.m. prioritization pass and meeting prep to meeting capture, end-of-day review, and clipping ingestion, all while keeping governed documents in Microsoft 365 where they belong.

They also dig into how a workflow matures from an on-demand skill to a scheduled skill, a loop, and finally a deterministic Python script that any interface can call. Scott shows off his locally served web dashboard pinned in Edge, Ben shares the custom Obsidian plugin Claude built for his command center, and they compare notes on choosing a source of truth (Markdown versus Asana), filtering Teams noise down to the signal, and what these systems actually cost to run each month.

Show Notes

A pattern for building personal knowledge bases using LLMs

LLM Wiki v2: extending the pattern with lessons from production

Agent Skills open standard

About agent skills for GitHub Copilot

Use Agent Skills in Visual Studio Code

Building effective agents

Conductor: Deterministic orchestration for multi-agent AI workflows

Obsidian Web Clipper

Prior podcast context