Start with the work, user, decision, or friction before choosing a model or tool.
WORKFLOW DESIGNAUTOMATIONEVALUATIONPROTOTYPING
AI LAB
Build it.
Test it.
Learn from it.
A working lab for practical AI, automation, product experiments, and systems thinking.
This is where I explore how AI can improve real work. I use modern models and development tools to frame problems, structure workflows, prototype ideas, test assumptions, and turn useful concepts into working systems.
pflugel-lab / active-session
$ lab --focus
Practical AI applied to real workflows$ method --current
Understand → structure → prototype → evaluate → improve$ control
Human review stays in the loopSeparate inputs, logic, actions, evidence, and human approvals so the system can be inspected.
Use AI-assisted development and automation to move from concept to something testable.
Look for failure modes, weak assumptions, poor outputs, and opportunities to simplify.
TOOL WORKBENCH
Tools are selected for the job, not the logo.
Filter the workbench, then select a tool to see where it fits in the way I work.
WORKFLOW LAB
AI is one part of the system.
Select a stage to see how the work moves from idea to controlled output.
01 / FRAME
Start with the real-world friction.
Define the user, the work, the decision, the constraint, and the outcome before deciding whether AI is even useful.
Questions I askWho is doing the work?What currently slows them down?What must remain under human control?
EXPERIMENT BOARD
Ideas become more useful when they have to survive contact with reality.
WORKING PROTOTYPE01
CrewClerks
Voice-first workflow design for service businesses. Natural language becomes structured customer, job, proposal, billing, reminder, and communication workflows.
Voice AIProduct DesignWorkflow LogicHuman Approval
Explore CrewClerks →
DECISION SYSTEM02
AI Opportunity Engine
Structured product and supplier research with hard gates, commercial scoring, confidence modeling, persistence, and n8n-oriented orchestration.
n8nLLMsScoringData Model
Explore Opportunity Engine →
WORKING BETA SOFTWARE03
FanDuel ATI Engine
Desktop software combining data workflows, simulation, lineup construction, late-slate controls, portfolio logic, and persistent learning.
PythonSimulationDesktop GUIIteration
View Product Walkthrough →
LIVE PORTFOLIO BUILD04
Pflugel.com
This site is itself part of the lab. It combines information architecture, responsive web development, interactive presentation, project storytelling, and AI-assisted iteration.
Web DevelopmentUXContent SystemsAI-Assisted Build
Return Home →
LAB PRINCIPLES
Useful beats impressive sounding.
Understand the work before reaching for AI.
Separate what is known from what is inferred.
Consequential actions need clear review points.
Test the idea while change is still inexpensive.
Bad outcomes are useful when they reveal what the system lacked.
A system is easier to improve when its rules can be seen and challenged.
THE POINT OF THE LAB
Learn enough to make the technology practical.
The goal is not to collect AI buzzwords. The goal is to understand what the tools can do, where they fail, and how to turn them into systems that help people work better.