Method and governance
The method every build follows, the chassis it runs on, the reference build you can walk through, and where this way of working has already run.
The method
A published, spec-driven method. Every stage produces a written artefact you can read: briefs, build-specs, decision records, ship records.
Where the load actually sits: the workflows, the systems of record, the people. An idea that does not make a person more capable is parked, in writing.
A build-spec with scope, constraints, architecture decisions, data model and an out-of-scope list. Agentic builds carry a governance sheet before any code: each agent's role, authority, bounds and hand-back conditions.
Fixed scope, tasks sized in days, decision records for every significant choice. You can read exactly what was decided and why.
Into your tenant with checklists as state: environment, secrets, rollback. A ship record notes what went live and what it costs to run.
The chassis
What every build stands on, and what happens to your data: it stays yours, in your tenant, in Australia East.
| Component | What it means for you |
|---|---|
| Azure subscription | You own the subscription and the data. ARK360 builds inside it and can be removed from it. |
| Entra ID | Identity and access are your directory, your policies, your joiners and leavers. |
| Data residency | Data resident in Australia East. Nothing routes through third-party services outside your tenancy. |
| Copilot Studio | The entry tier for governed assistants inside Microsoft 365, where your teams already work. |
| Claude Agent SDK via Azure AI Foundry | Deeper agentic builds authenticate through Azure AI Foundry, so billing and governance stay in your tenant. |
| Reusable MCP integrations | Connections to your systems of record are built once as governed integrations and reused across builds. |
| Outbound actions | Nothing outbound is ever autonomous. Drafts are prepared for a person; a person sends. |
Governance
None of this is compliance theatre. It is the reason the build still behaves in month six: measured, bounded, auditable, and priced to run.
Every build ships with evaluations that run before and after changes, so behaviour is measured, not assumed.
Workflows are designed around the person: the system prepares, the person decides. Hand-back conditions are written down before a build starts.
Every agent action lands in an audit record: what acted, under which authority, with what outcome and cost.
Each agent holds the narrowest authority that does the job, recorded in a governance sheet you can read.
Token and compute spend is capped and visible per workflow, so the running cost is a number, not a surprise.
What it is
It has no paying customers and no live operators. It is a reference build and an integration substrate, not a product I am selling you. It exists so you can see the method working before you commission anything.
That status is stated plainly because the value of this page depends on it. Anyone can show you a demo. A complete platform, built under the same method offered to you, is something you can interrogate: how it is structured, how authority is bounded, what it costs to run.
The walkthrough
Each part exists because a build-to-rent operator needs it on day one. Together they cover the operational surface end to end.
The permission and credential engine: who can open which door, when, under whose authority, and the auditable event stream every decision leaves behind.
A .NET MAUI app for residents: access, requests and communications in one place, against the same core the operator sees.
A Blazor console for the operating team: occupancy, access events, requests and the day's work in one view.
Physical access hardware treated as a governed boundary: hardware events flow in, permission decisions flow out, and both are logged.
Estate configuration, roles and the reporting an operator answers for, drawn from the same records the workflows write.
Built on .NET Aspire, ASP.NET Core, Blazor and .NET MAUI, with MCP integrations.
Why it matters to you
The useful part for a buyer is not the platform itself. It is what sits underneath: governed integration patterns for identity, access events, resident communications and reporting, built once and reusable. Client builds draw on that substrate instead of starting from a blank page.
If you commission a build, you are not buying BTR OS. You are engaging the method that produced it, applied to your workflows, your systems of record and your tenant.
Read the method behind itDelivery record
Federal government
Delivered as employed work, not ARK360 client work.
Stood up the governance, pipelines and maker enablement for a federal agency's Power Platform estate: environment strategy, DLP policy, application lifecycle and support model.
What persisted: The CoE operating model and governance remained in daily use after handover.
Civil construction
Delivered as contract work, not ARK360 client work.
Designed and delivered agentic AI workflows for engineering and commercial teams on the Microsoft platform, including a technical health review agent that produced its first working proof of concept in one week and moved into a production build.
What persisted: The governed-agent pattern and the delivery method carried into subsequent internal builds.
High assurance
Delivered as contract work, not ARK360 client work.
Designed an Azure OpenAI integration pattern for a high-assurance environment: private networking, identity-bound access, prompt and output logging, and data-residency controls in the customer tenant.
What persisted: The pattern was adopted as the reference for later AI workloads in that environment.
Property and development
Delivered as employed and contract work, not ARK360 client work.
Seventeen-plus years across property, development and construction businesses on the Microsoft platform: from SharePoint and .NET estates through Power Platform to AI-era workloads, always inside the owner's tenant.
What persisted: Platforms and operating patterns that teams still run.
Names are withheld deliberately. Employers and clients are described by category, and outcomes are limited to what I can state plainly and stand behind. Specifics can be discussed in conversation where confidentiality allows.