| name | apply-frameworks |
| description | Apply Brian Madden's analytical frameworks to new situations, questions, or business problems. Brian's frameworks are mental models for understanding AI's transformation of knowledge work and the enterprise. Use when the user wants to analyze a situation through Brian's lens, understand an AI trend, or get a structured perspective on enterprise technology. Trigger with "apply Brian's framework to...", "how would Brian analyze...", "what does the 80/20 framework say about...", or "use Brian's thinking to evaluate...". |
Apply Frameworks
Brian Madden has developed several interconnected frameworks for understanding how AI transforms knowledge work. These aren't academic models — they come from 32 years of watching enterprise technology waves and pattern-matching how change actually happens in large organizations.
Core frameworks
Use get_framework from the brianmadden-ai MCP server to load any of these:
The invisible 80% (invisible-80-percent)
The most important work in knowledge work is tacit — it lives in workers' heads, not in apps or documents. Corporate AI strategies focus on the visible 20% (structured workflows, documented processes) while missing the 80% that actually drives value. Worker-led AI adoption is the only way to reach this invisible majority.
Factory electrification (factory-electrification)
The 1880s-1920s transition from steam to electric power is the best historical analogy for AI transformation. Key lesson: the technology was available for 40 years before factories were redesigned to exploit it. The real transformation required rethinking the entire production process, not just swapping power sources. Same pattern applies to AI — bolting AI onto existing workflows misses the point.
Post-application era (post-application-era)
Applications are middleware between humans and data/services. AI doesn't need that middleware. We're entering an era where AI agents interact directly with APIs, data, and services — the application layer dissolves. This has massive implications for enterprise software, governance, and how work gets done.
Workspace as control plane (workspace-as-control-plane)
As AI agents proliferate, the enterprise workspace becomes the governance and orchestration layer — the control plane through which human-AI collaboration is managed, secured, and observed. Not the tool where work happens, but the infrastructure that makes safe work possible.
The bitter lesson (bitter-lesson)
Stop engineering, start enabling. The invisible 80% of knowledge work isn't something to capture — it's human legacy infrastructure that AI doesn't need. AI routes from inputs to outputs without traversing human cognitive scaffolding. The bitter lesson is the meta-principle underneath multiple other frameworks.
Delegation not automation (delegation-not-automation)
Workers delegate to AI, they don't automate with it. The skills hierarchy: worker → cognitive extension → skills → agentic sub-processes → workflow automation. The industry invests at the bottom (automation); the value is at the top (delegation). Agents are a form factor, not a solution.
Five levels of AI in knowledge work (five-levels-of-ai-in-knowledge-work)
From spicy search engine to dark knowledge factory. Adapted from Dan Shapiro's five levels of AI coding assistance — the coding-as-leading-indicator analysis applied to knowledge work. As AI takes on more production work, the human role shifts from doing to directing to verifying. Most enterprise knowledge workers are at Level 2-3 thinking they've maxed out. Largely supersedes the 7-stage roadmap: the 7-stage was mechanical (what AI can do), the five levels map the human experience (what the worker's role becomes).
7-stage roadmap (7-stage-roadmap)
A progression from AI-as-assistant to AI-as-autonomous-colleague. Most organizations are at stages 1-2 (individual productivity tools). The interesting work is in stages 4-7 where AI becomes a genuine collaborator with its own context, memory, and agency. Largely superseded by the five-levels framework, which maps the human experience rather than the mechanical capability.
Subscribable brains (subscribable-brains)
Experts should stop publishing content and start publishing their structured knowledge repos. Subscribers sync via git and MCP, integrating the expert's knowledge into their own AI systems. The creator's maintenance of their own second brain is the product.
How to apply frameworks
When analyzing a situation through Brian's lens:
-
Identify which frameworks are relevant. Most situations touch 2-3 frameworks. The invisible 80% and factory electrification are almost always relevant to enterprise AI discussions.
-
Load the full framework. Use get_framework to get the complete version with evidence and nuance, not just the summary above.
-
Apply with specificity. Don't just name-drop the framework. Show how it reframes the specific situation. Brian's signature move is showing why the conventional framing misses the point.
-
Acknowledge the connections. The frameworks are interconnected. The invisible 80% explains why factory electrification is hard. The post-application era explains what the workspace-as-control-plane needs to govern. Subscribable brains are what personal AI knowledge systems look like at scale.
-
Ground in practitioner reality. Brian thinks like someone who's watched enterprises adopt technology for three decades. The vendor narrative is always cleaner than reality. Change is messier, slower, and more worker-driven than the slides suggest.
What Brian's frameworks are NOT
- They're not vendor positioning (even though Brian works at Citrix)
- They're not predictive timelines ("AI will do X by 2027")
- They're not prescriptive playbooks ("do these 5 steps")
- They ARE lenses for understanding what's actually happening vs. what the hype cycle claims
