Repeatable deployment patterns for AI enablement engineers working in regulated enterprise contexts. Field-tested. Eval-validated. Sanitized for public consumption.
I'm Logan Freeman. I run AI enablement engagements at enterprise scale (currently designing AI learning architecture for a large regulated utility, roughly 120,000 employees across multiple operating companies) and I build governance infrastructure for agentic AI systems (WESUMN, at wesummonit.com). This repo is where I publish the patterns I actually use, the failure modes I've actually hit, and the artifacts I'd hand a customer on day one of a new engagement.
If you're new to this work, read Pattern 01: The Four-Stage Engagement Pipeline first. It's the operating system underneath every other pattern in this repo.
If you're navigating an EU AI Act or Colorado SB-205 conversation right now, jump to Pattern 02: Translating Regulatory Requirements into Product Features.
Four patterns are complete and readable end to end. The rest are scoped and in progress, marked honestly below so you know what you're clicking into.
| # | Pattern | Status | Tags |
|---|---|---|---|
| 01 | Four-Stage Engagement Pipeline | Complete | engagement, process, enterprise |
| 02 | Translating Regulatory Requirements into Product Features | Complete | regulated-domain, eu-ai-act, compliance |
| 09 | The 8-Pillar AI Platform Enablement Playbook | Complete | strategy, governance, measurement |
| 10 | Ask / Delegate / Orchestrate: an AI Fluency Ladder | Complete | fluency, adoption, individual |
| 03 | A-Team / B-Team Multi-Agent Review Protocol | In progress | multi-agent, quality-gate, evals |
| 04 | Spec-Driven Development with AI Partners | In progress | claude-code, context-engineering, workflow |
| 05 | The AI Use Case Register | In progress | governance, intake, assessment |
| 06 | Cognitive Load Protections in Agent UX | In progress | agent-ux, accessibility, production |
| 07 | AuDHD-First Defaults as Structural Constraint | In progress | design-philosophy, accessibility, meta |
| 08 | Self-Serve Documentation Hubs | In progress | handoff, adoption, measurement |
Most AI enablement writing in 2026 is either marketing prose or vendor cookbooks. Neither helps you when you're three weeks into a regulated enterprise deployment and the legal team has just asked what Article 14 actually requires of your agent.
This repo is what I wish existed when I started doing this work. The patterns are named, the failure modes are explicit, and every pattern that can carry a runnable artifact has one. Field reports are sanitized but real.
If you're hiring for AI Enablement, Forward Deployed Engineering, or Developer Advocacy roles and you want to see how I think, the patterns are the answer. The contact info is at the bottom.
Each pattern follows the same shape: a problem statement, a "why this exists" anchor (about 200 words), an architecture or workflow diagram, a runnable artifact (template or code), an eval section showing the pattern wins on some measurable axis, named failure modes, and a field report sidebar drawn from real engagements.
If you want the supporting infrastructure, see:
anti-patterns/, short essays on what fails and whyevals/, calibrated judges for the patterns that admit oneincidents/, sanitized postmortems from real engagementsrunbooks/, operational checklists, the Handoff artifactsmcp-servers/, real MCP servers wiring Notion and Asana into agentsfield-reports/, long-form writeups of engagements
ljfreeman83@gmail.com · linkedin.com/in/loganfreemanai · straysouth.com
If your team is hiring for AI Enablement, Forward Deployed Engineering, Developer Advocacy, or AI Product Management roles in 2026, my inbox is open.
Patterns, templates, and prose are CC BY 4.0. Code samples are Apache 2.0 unless otherwise noted.