I ship AI systems into production, and tell you when not to build one.

I'm Adam Broons.

I run an independent practice taking AI systems from decision into production, and then getting organisations to actually use them.

I am CTO at OopZee and founder of Scorafy, and I consult through Cognitiv. What I have built and delivered: OopZee, a vehicle-damage reporting app live on the web and released through store review on the Apple App Store and Google Play; Scorafy, a multi-tenant assessment platform running in two regions, with tenant isolation enforced at the database layer, EU and AU data residency, a published data processing agreement, multi-factor authentication and a full audit trail; the full digital platform for a Danish women's football organisation, including the match-day systems used at its first live event; and LeagueKit, a tabletop campaign platform with two thousand automated tests behind it.

The work I am proudest of is a job I talked a client out of. A 350-staff hospitality operator hired me to build AI rostering. I assessed it, concluded a custom build was not warranted, and recommended and brokered an established platform instead. They got a working system and never paid for software they would have had to maintain.

Most AI programmes fail in one of two places: nobody adopts the tools, or nobody can say whether the output is good enough to rely on. I work on both, and the second is where I am unusual. Before technology I spent five years teaching and marking inside Western Australia's externally moderated examination system - designing marking keys, moderating against other assessors, defending grade decisions under formal review. Assessing work against a defined standard, and proving the standard held, was the job.

Before that: small-business banking at ANZ, business development for an internet company through its pre-float raise, and running my own venue. That breadth is why I can sit with a non-technical operator, understand the business, and turn it into software people use.

I direct AI development tools and I am accountable for what ships. I set the architecture and the policy constraints, review and reject output, and own the result in production. I do not hand-write the implementation and I do not claim depth I do not have.