Hang Ten Systems

Software built around the business

Hang Ten combines an enterprise AI platform called Hobie with engineers who work alongside customer teams. It offers software development and modernisation, with pricing agreed against defined outcomes.

Hobie’s published architecture includes a shared model of business terms and relationships, queries against existing systems, and an audit layer. It describes turning structured requests into read-only queries, with access tied to the caller and answers traced to their sources.

The technical test

I would start with a business term that sounds simple, such as revenue. Different teams may define it differently. A useful semantic layer needs to make that definition explicit before an agent can answer consistently.

Read-only queries reduce one kind of risk, but they can still reveal information to the wrong person. I would test access boundaries and whether the answer cites the correct version of the underlying logic.

For software delivery, I would examine the whole release. Generated code needs tests, integration, deployment and a clear owner when something breaks. Faster coding helps most when those surrounding steps are dependable.

A services model can be attractive

I would not dismiss a business simply because people are involved in delivery. Customers can value a team that takes responsibility for a difficult outcome. The question is how much useful work that team can repeat.

Outcome-based pricing makes the definition of success important. Clear scope and acceptance criteria can align both sides. Unclear requirements can turn a fixed outcome into unlimited work.

The financial lens I would use is delivery effort per completed project, the share of work that can be reused and the cost of supporting it afterwards.

The moat and my take

I see a possible moat in reusable software, domain knowledge and customer trust reinforcing one another. The strongest version gets better with each deployment while protecting each customer’s data.

I like the focus on the unglamorous work required to make AI useful inside a company. I would judge the direction by whether delivery becomes more repeatable, with less reinvention each time. That is where a good services business could build a lasting technical advantage.