Fixing an AI coding mistake addresses one incident. The Ratchet Loop goes further by diagnosing why it survived and turning the lesson into a durable...
Read More
Alex Solomon | Business Systems Architect
I design systems that make complex work reliable.
My work sits at the intersection of business operations, software, and governed AI. I design systems that connect decisions, evidence, people, and workflows into dependable execution.
Based in southwest Missouri. Building at the intersection of operations, software, and human judgment.
What I work on.
Building systems around the work, not just tools for the work.
Much of my current work centers on a practical question:
How can businesses use increasingly capable software and AI agents without surrendering evidence, accountability, or human authority?
Governed AI Systems
Company-owned software that places AI inside defined evidence, permission, verification, and approval boundaries.
These systems are designed to help AI perform meaningful work without quietly becoming the authority over the work itself.
Business and Operational Systems
The processes, information architecture, ownership rules, and workflows behind reliable execution.
This includes defining sources of truth, clarifying how decisions are made, and turning fragmented work into a coherent operating system.
Marketing Operations
CRM, research, strategy, production, reporting, automation, and performance data brought together into one connected environment.
Marketing operations has been the primary proving ground for much of my systems work.
Ratchet Governed Agentic Development
Inspired by established ideas in software quality ratchets, regression testing, and agent harness engineering, Ratchet Development explores how verified failures can strengthen the environment future work inherits.
A failure might become a regression test. It might also become a typed contract, permission boundary, database constraint, provenance requirement, or approval gate.
The goal is not to eliminate every mistake or create more process around every decision. It is to understand why meaningful failures survived and decide whether a proportionate, durable control can make them harder to repeat.
The model proposes. Deterministic systems verify. Humans retain authority.
Selected work.
Turning complex operations into reliable systems.
Governed Marketing Operating System
CRM and Operational Architecture
Creative Systems and Service Design
Governed Marketing Operating System
A company-owned platform designed to connect research, evidence, strategy, production, approvals, and reporting inside one governed operating environment.
Focus areas
- Evidence and provenance
- Human approvals
- Deterministic verification
- Pipeline execution
- Permissions and audit history
CRM and Operational Architecture
Redesigning the way business information moves through teams, systems, and customer relationships.
The work often includes data models, lifecycle definitions, automation, ownership rules, reporting, adoption, and the handoffs between sales, marketing, delivery, and leadership.
Focus areas
- HubSpot architecture
- Sources of truth
- Automation and workflows
- Reporting systems
- Operational alignment
Creative Systems and Service Design
Before my work centered on business and software systems, I spent years building creative teams, service lines, websites, brands, video operations, and repeatable production processes.
That experience continues to shape how I approach usability, communication, modularity, and the people responsible for operating a system.
Focus areas
- Design systems
- Service development
- Creative operations
- Team workflows
- Digital experiences
Writing from the work.
I write about the systems, mistakes, and operating principles emerging from my work with business operations, software, and AI agents.
Claude Code Builds. Codex Reviews. I Retain Authority.
I use Claude Code to build and OpenAI Codex to review, but I keep a hard wall between implementation and independent critique. Neither model gets...
Read More
Autonomy Is Not Authority: The Human Role in Agentic Software Development
Agentic software development works best when AI agents have meaningful implementation autonomy, deterministic systems verify their work, and humans retain authority over meaning, architecture, and...
Read More
About me.
I was a builder long before I started building software.
I grew up helping my dad with construction projects, wanting to become an architect, and making anything I could figure out how to make.
That instinct led me through graphic design, photography, filmmaking, websites, creative leadership, marketing operations, and eventually software and governed AI.
The tools and scale have changed. The thread connecting the work has not: I like understanding how complex things are built, how their parts relate, and how the entire system can work better.
Let’s build a better system.
I am always interested in thoughtful conversations about governed AI, business systems, operational design, and the future of work with increasingly capable software.