The Agent Operator Framework
The most valuable person in AI isn't the one who builds it. It's the one who runs it.
A living framework for directing AI agent systems — for individuals, teams, and companies. Built from real work, updated as AI evolves.
See how it works ↓This is a framework, not a product
AI is not a one-time thing to learn. It keeps changing — new tools, new capabilities, new ways of working. The Agent Operator Framework isn't a static course you finish once. It's a way of thinking about directing AI systems that we keep testing, updating, and rebuilding as the technology moves. What you learn here holds up next year because we keep it current, not because it was frozen the day you bought it.
The framework itself
Every AI agent system — no matter the tools — runs on four components: Brain, Nervous System, Memory, Voice. Directing them well takes six skills: prompt engineering, workflow logic, data literacy, output auditing, escalation judgment, and domain fluency. This is the architecture underneath everything else on this site.
Brain
The reasoning engine — like Claude — that interprets instructions and makes decisions. This is what you're directing.
Nervous System
The connections and workflows that let the brain act across your other tools — the wiring between thought and action.
Memory
What the system remembers between runs. Good memory compounds intelligence over time; no memory resets it every time.
Voice
How the system talks back — to you, your team, your customers. The interface where trust is won or lost.
The Six Skills
Prompt Engineering
Writing AI instructions that reliably produce what you need. The skill everything else depends on.
CriticalWorkflow Logic
Reading automated processes, tracing failures, describing fixes — without writing code.
HighData Literacy
Finding, reading, and correcting the data your agents use. Bad fuel means bad output.
HighOutput Auditing
Systematically reviewing agent outputs for quality, accuracy, brand, and compliance.
EssentialEscalation Judgment
Knowing when to stop the automation and bring in a human. The most underrated skill in the profile.
CriticalDomain Fluency
Role-specific competencies for how agents handle your field — marketing, finance, legal, ops, healthcare.
Role-specificWays to use this framework
The Agent Operator Guide
A self-paced walkthrough of the framework — seven chapters, a 90-day path, instant PDF download. $57.
Get the guide →The Cohort
Nine live sessions, real work, a community, and certification. $697.
See the next cohort →The Community
Two live sessions a month, plus a private Slack group, where cohort graduates keep building — the ongoing practice space after the course. $199/month.
Explore the community →Smart Edge Technologies
Have your team's AI systems designed and implemented for you, built on this same framework.
Talk to Smart Edge →What people have built with it
Testimonials
"I built a capability one-pager for my business and got real UI/UX feedback using Claude's Chrome extension — genuinely useful, immediate results. But the bigger shift was in how I think about Claude: I went from treating it like a single thinking partner to running it like three skilled employees, each with their own output to leverage."
"I've been deep in the AI space for about a year, using it to transform how I work and helping others see its real operational value. But this program cracked something open for me — it showed me how to take everything I'd been doing and actually wire it into the functioning of my business, not just as a tool I use, but as the engine that runs it. That's a completely different level."
"Before this program I was building with Claude in ways that worked, but self-taught, without a shared vocabulary for what I was doing. This cohort gave me a framework to name it — and confirmed that the most important design decision in AI adoption isn't which model you choose, it's how you structure the relationship between human judgment and the system."
Built by an operator, not a commentator
The framework comes from Christina Dills — founder of Smart Edge Technologies, 13 years at Microsoft, credentials from Harvard, Columbia, and IBM — who runs three AI-powered companies on the exact systems it describes. Everything on this site comes from real deployments: what worked, what failed, and what it takes to stay in control.
This keeps evolving
New Claude features, new tools, new ways of working — we update the framework as they land, and cohort members get those updates first.