Design the system around the problem
Systems are shaped by the type of problem—not by tools or stack. Growth, information, product, and decision systems require different structures, workflows, decision points, and definitions of success.
← Home
Build
Build the system.
Turn a framed problem and allocation into a system that actually works in practice.
What makes a system actually work
Building is not just producing a tool. The system has to work in use — under real conditions, over time, and as it evolves.
Control decision load
A system should reduce unnecessary decisions without eliminating judgment. Too many decisions slow the work down. Too few decisions make the system blind.
Build small, but scalable
The first version should be small enough to test, but structured so it can be extended without rebuilding from scratch.
Reveal problems early
A good system makes it visible where work gets stuck, where outputs degrade, and where assumptions break.
Integrate testing into the build
Testing is not a separate phase. Usage, outputs, and failure points should be observable as part of the system itself.
Hold under real use
A system is only real if it works with real users, real constraints, time pressure, and incomplete information.
What I build
Systems for different kinds of problems
Different problems require different systems. The common thread is that the system is designed around the problem, not around a preferred tool.
Growth Systems
Outbound workflows, lead pipelines, SEO systems, content workflows, and acquisition systems.
Intelligence Systems
Text analysis, communication analysis, internal knowledge systems, and pattern-detection workflows.
Product Systems
MVPs, prototypes, web applications, APIs, internal tools, and business platforms.
Decision Systems
Decision support, concept refinement, interpretation systems, and human-AI workflow design.
The full approach
1
Frame the problem
Clarify what is actually being solved before anything is built.
2
Allocate intelligence
Decide what belongs to software, AI, and human judgment.
3
Build the system
Design and implement a system that works in practice.
Most systems underperform because at least one of these steps is wrong.
Have a system worth building?
Let’s frame the problem, allocate the right intelligence, and build the system that fits.