I spent thirteen years building my first group of companies and exited in 2020. Since then I have built a portfolio of software products, each in operation in its own market.
I remember building an automation for the project folders in my own company.
Every new job entered our CRM with a job number. That event created the folder structure on the server, with the same naming each time. It was a small piece of software doing a defined job in the business I ran.
I liked the simplicity of it. The record we already relied on became the trigger for the next piece of work. People could get on with the job and use the files where they expected to find them.
That experience still shapes how I approach software. I start with the process, including the record that tells me what has happened and what needs to happen next. Once that is clear, I can judge which work a system should do and where a person needs to decide.
Repeating a task makes it a candidate for automation. Measuring the time it takes gives me a basis for deciding whether the investment is worthwhile. I want that baseline before the change is designed, so the result can be compared with what came before.
I also want the people using the system involved in testing it. A working trial reveals gaps that a specification alone will miss. It gives the business evidence to use when deciding whether to make the change permanent.
That is the thinking behind Seichō. I want the commercial recommendation to carry through into the software and the way it runs. The owner should know which record to trust, who can approve a change and what happens when the system fails.
AI belongs in that same discipline. I use automation for work that follows defined rules. Where a model adds value, its role needs clear limits and an accountable person to approve consequential output.
The folder automation gave me a useful standard to work from: begin with a real task and make its execution dependable. Its value rests on a measured change in how the work runs. I expect that standard to hold as the scope grows.
Connor McAuley
Principal and Chief Systems Architect