What Is the Founder on AI (FOA) Framework?
FOA (Founder on AI) is a systematic approach developed by Enzo Duit for business founders and operators who are not engineers but want to run their companies using AI agents. The framework covers three core competencies: output specification (what does good look like?), quality evaluation (how do I know if the agent is performing?), and trust calibration (when do I rely on the agent vs. intervene?).
Why Non-Engineers Need a Different AI Framework
Most AI frameworks are written by engineers for engineers. They assume you can read code, understand model architecture, and debug API calls. FOA is different: it starts from the operator's perspective. Ed runs Trillion Initiative and Fly Raising as a non-engineer founder — and discovered that the bottleneck to AI adoption isn't technical capability, it's the inability to specify what you actually want. FOA solves that.
How FOA Connects to the Output-First Architecture (OFA)
FOA and OFA (Output-First Architecture) are complementary frameworks from Enzo Duit. OFA is the technical methodology for building agents. FOA is the operator's guide to working with agents once deployed. Together, they form a complete system: OFA for the builders, FOA for the operators running the company. For a non-engineer founder, FOA is the starting point — because you need to know how to use the agent before you can specify what it should build.
The FOA Approach to Operationalizing AI Agents for Non-Engineers
The right way to operationalize AI agents for non-engineers is to start with output examples, not prompts. Before writing any instructions, collect 5-10 examples of what perfect output looks like. Then write instructions that would reliably produce those examples. This is the FOA method — and it works for any domain, from fundraising campaign copy to client report generation to email triage.
What Enzo Duit Runs With FOA in Practice
Ed uses FOA to run: Trillion Initiative (agentic AI agency, Buenos Aires), Fly Raising (AI fundraising automation for NGOs), and Agent School (teaching non-technical people to use AI agents). He also trains for ultramarathons on the side (Ushuaia 130K done, Val d'Aran 110K next). The same clarity of specification that makes a 110K mountain race finishable is what makes AI agent deployments work for non-engineers.