AI-Native Organisations Run on Skills: How to Structure and Scale Them — Imad Touil, QuantumBlack
Summary
The talk argues that in AI-native organizations, organizational know-how resides primarily within 'skills.' While skills offer deterministic outcomes for complex workflows, ungoverned skills quickly accumulate as a form of technical debt due to duplication and quality decay. To scale effectively, organizations must adopt governance principles—borrowing from the microservices era—by implementing a centralized skills platform that provides metadata, search, versioning, access control, and clear ownership.
Key takeaways
-
Skills are the core of organizational know-how
11:43
The agentic software stack relies heavily on 'skills' (the deterministic component) within its outer workflow loop. If skills are unstructured, the resulting workflow is not truly deterministic.
-
Governance prevents technical debt
20:30
Without governance, teams repeatedly build similar skills (duplication), quality degrades because skills aren't retested against new models, and ownership is unclear. This creates a new class of technical debt.
-
A centralized platform is mandatory for scale
Scaling requires a central skills catalog with metadata, dependency mapping, versioning (to pull the current release), access control, and named human owners across architecture, infrastructure, and security.
Technical details
-
Agentic Software Stack Architecture
133s
The stack has two loops: the inner loop (code/agents harness) includes components like context manager, tools, MCPs, memories, states, and skills loader. The outer loop (workflows) utilizes skills, sub-agents, MCP servers, and hooks. Enabling components include an environment sandbox, an MCP gateway, a model gateway (for managing LMS), a graph knowledge graph (abstracting core systems/skills registry), and a workflow marketplace.
-
Skill Design Principles
419s
Skills must adhere to microservices design principles: they need to be reusable, modular, discoverable (portable across different harnesses/workflows), specialized (defining one specific task), and composable. Consistency and deterministic outcomes are key.
-
Workflow Lifecycle Complexity
256s
The end-to-end product lifecycle is far more complex than the simple 'Specify -> Plan -> Task -> Implement' model. It includes defining product strategy, market research, competitive analysis, discovery (problem statements/user stories), data preparation/cataloging, platform engineering ops, and continuous optimization.
Mentioned resources
- Skills Registry
- MCP Gateway
Channel & topics
Watch on YouTube · Back to latest
This independent, AI-assisted summary is provided for commentary and informational purposes. It may contain errors or omit important context. Please watch the original video for the creator's complete presentation. Video, thumbnail, and related copyrights belong to their respective owners.