AI-Native Organisations Run on Skills: How to Structure and Scale Them — Imad Touil, QuantumBlack
Source
youtube.com
Author
AI Engineer
Date
Context
Reusable AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → skills become a maintenance responsibility when several teams depend on them. A agent skillA reusable instruction file that teaches an agent how to do one job well — the procedure, the tools, and what counts as done.Full definition → packages instructions, scripts and supporting resources, as Anthropic explains in its design article. QuantumBlack's conference talk focuses on what happens after people begin sharing those packages: duplicated work, unclear ownership and quality that can deteriorate as models change.
Building on the skill format covered earlier, the talk proposes a searchable catalog with named owners and versioned dependencies. Teams would know which release they use and who must retest it. Because skills can execute scripts, reviewing an imported package also becomes part of managing software dependencies.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
agent skill — A reusable instruction file that teaches an agent how to do one job well — the procedure, the tools, and what counts as done.
MCP — The Model Context Protocol — an open standard that lets any AI assistant plug into any tool or data source without custom integration code.
videoDeploying AI Inside a Bank Is Not an Engineering Problem — Paden Gayle, Applied AI Engineers
videoThe Agent Works. Does Anyone Trust It? — Michele Grant, Leading at Scale
videoWhat Technical Due Diligence Finds When the Deck Says “70% Automated” — Soukaïna Abelhad
videoTokens Are the Cheap Part — Steve Jamison, Jellyfish
