The methodologies I have developed and the infrastructure that proves them.
The writing on this site argues that AI transformation is fundamentally an organizational problem; that specification quality is a leadership artifact; and that the human in the room becomes more valuable as the technology becomes more capable. The methodologies below name the instruments I use to make those arguments operational. The implementations below are what those instruments produce when applied to specific problems.
Seven frameworks for the VP or Director running an AI transformation. Each framework addresses a specific failure mode that kills transformations at the leadership layer.
Framework 1
AI systems reflecting vendor defaults instead of your culture. When you don't specify what your AI should reflect, vendor defaults fill the gap.
Read the framework →Framework 2
AI doesn't preserve the pyramid. The diamond emerged as the consensus answer — then PwC walked it back. The post-pyramid shape is in flux. But whatever shape you bet on, you're betting on a working judgment pipeline. The shape question is only half the problem.
Read the framework →Framework 3
Leadership teams appearing aligned while holding different assumptions. Everyone agrees they're "all in on AI." Nobody agrees on what that means.
Read the framework →Framework 4
Adoption resistance misread as change resistance. People aren't resisting tools. They're protecting professional identities that the transformation hasn't offered a replacement path for.
Read the framework →Framework 5
Governance documents too rigid to handle novel situations. A compliance document fails exactly when governance matters most: when something new comes up.
Read the framework →Framework 6
The Valley misread as failure. Month-3-through-5 friction is predictable and necessary. Leaders who don't name it before it arrives lose credibility when it appears on schedule.
Read the framework →Framework 7
Measurement systems tracking effort instead of capability development. High usage does not equal good outcomes. Some of the worst AI use looks great on usage dashboards.
Read the framework →Named instruments for measuring the gap between stated policy and lived user experience. Each methodology has a permanent home and links to the implementations that prove it.
Shipped projects that demonstrate the methodologies in practice. Open-source where possible, revocation-resistant by design, free for the people who need them.
Implementation
A free, open-source directory of subscription cancellation instructions for more than two hundred services. Each entry carries a Friction Score and documents the dark patterns the company uses to retain customers. No accounts, no tracking, no data collection. The first public implementation of the Friction Score methodology.
Visit Cancel Freely →Implementation
A companion directory to Cancel Freely for personal data requests. Each company entry scores the friction a user actually encounters when exercising rights under GDPR, CCPA, CPRA, and other privacy regimes, alongside copy-ready request templates and jurisdiction-specific rights guides for the EU, California, and US states with active privacy laws. Includes a walkthrough of California's DROP portal. Open source, no accounts, no tracking.
Visit DeleteFreely →Implementation
A guide to exiting Big Tech ecosystems without losing the capabilities they provide. Seventy-plus pages on alternatives, migration paths, and the trade-offs nobody else explains honestly. Built on the conviction that the best privacy posture is the one a non-technical user can actually adopt.
Visit Ditch the Mega →Implementation
A bereavement resource for closing the digital accounts of someone who has died. Twenty-nine-plus service-specific guides covering what to gather, what to expect, and how to push through the friction companies create at the worst possible moment. Built after my mother's unexpected death in 2007 left a problem nobody had documented.
Visit Closing Accounts →Implementation
A digital exhibition of forty-two fictional institutional websites produced through human-AI collaboration. The subject of a published case study on where human judgment is irreplaceable in AI-assisted creative work. The most distinctive demonstration in the practice of what spec-driven collaboration can produce when the human keeps the decisions that matter.
Visit Peregrin Arc →The writing that develops these methodologies and reflects on these implementations lives on Signal.
Paid tools
All 7 frameworks plus worksheets, session guides, a 12-month roadmap, and 11 AI prompts. Built for the VP or Director running an AI transformation.
Get the Toolkit ($97) →The diagnostic framework for identifying and working with High-Will/Low-Skill organizations. The signals, the qualifying conversation, and the five moves that work.
Get the Field Guide ($47) →