Learning loop
Business-facing agents
Learning the platforms businesses are likely to buy while staying grounded in workflow, controls, and practical adoption.
Mash Prime · Perth · builder/operator
I build and test AI agent systems, release the useful pieces, and apply the lessons to business workflows where the pain is real.
The current work is model-agnostic at the core: agents, memory, approvals, tools, dashboards, and workflows. The business-facing learning loop matters because that is where adoption is going, but the deeper goal is useful systems that survive outside one model or vendor.
what this site is now
A living hub for business problems, agent experiments, writing, diagrams, and lessons worth reusing.
Learning loop
Learning the platforms businesses are likely to buy while staying grounded in workflow, controls, and practical adoption.
Commercial loop
Working on the boring, expensive edge of business systems: orders, reporting, approvals, admin handoffs, and dashboard glue.
Public loop
Turning actual build work into writing, examples, and reusable patterns instead of content theatre.
operating loop
The site should make the work legible: what I am learning, what I am building, what is ready to use, and what can help a business now.
Find real friction
Build a small working system
Review what survived
Teach the useful pattern
public doors
workflows / consulting
Agentic workflow consulting for Australian businesses: automation, dashboards, approvals, and business-system glue.
media company / creative operations
Bass music label and media company. A live operating ground for release workflows, publishing systems, and creative infrastructure.
public notes
Short essays and build notes on agentic systems, practical workflows, memory, approvals, and what survives real use.
work with me
The commercial lane is not “AI magic.” It is reducing manual admin around orders, reporting, approvals, dashboards, and business systems without replacing everything at once.
See the work page →learn with me
I’m documenting the process as I learn business-facing agent platforms, compare them to what I built by hand, and separate useful workflow patterns from vendor hype.
Follow the learning log →system log
A working note on what six months of building with AI agents has taught me: the scarce skill is not magic prompts, it is managing work.
The companion to my Claude Community lightning talk in Perth. The terminal is not as scary as it looks, and a single folder can climb a surprisingly long way.
From a job title held by women doing calculations, to room-sized machines, to the personal computer, and now to always-on agents. A reflection on where computing is heading and what to build for it.
next useful action