Outcome before novelty
Use AI or new technology because it improves the work, not because it is fashionable.
About
One partner who can see the business and the system.
Steven Esau brings accounting, economics, operational thinking, and software development together to solve problems in context.
Business systems · AI automation
The useful intersection
I am interested in what happens between the spreadsheet, the policy, the customer, and the person doing the work.
That is usually where delays, duplicated effort, and missing information live. It is also where a well-designed system can make an immediate difference.
Through Tonga Market Ltd, I have developed marketplaces, finance websites, education and ERP concepts, project tools, asset applications, and AI-assisted workflows. The technology changes; the principle does not: understand the work, protect the people and information involved, then build for a measurable outcome.
Being based in Tonga gives this work a practical perspective. Solutions need to respect local realities - cost, connectivity, skills, support, and trust - while still meeting a modern technical standard.
Working principles
The standards behind the build.
Privacy is structural
Protect client information through architecture, access control, and disciplined publishing.
Show the working truth
Prototype early enough that users can test the process rather than approve an abstract promise.
Keep people accountable
Automation should support judgment and make responsibility clearer, not make it disappear.
Tonga Market Ltd
Built in Tonga, with an expanding view of what local organisations can create.
Tonga Market Ltd began with business listings and digital commerce. Its work now extends into custom business applications, integration, and responsible AI automation while retaining the original goal of making useful technology more accessible.
Publishing guardrails
Credible proof without exposing private work.
Public pages can describe services, capabilities, and already-public demonstrations. They must not expose credentials, unpublished proposals, private client records, production data, or security details that would help an attacker.
Private prototype opportunities remain private unless written approval is given to publish them.
A useful first conversation
Bring the process that is costing you time.
We will map the problem, identify the quickest valuable improvement, and tell you honestly what should happen next.