// Services

Four kinds of work.

Services aren’t the starting point. Discovery is.

// Discovery first

Services aren’t where this starts. Discovery is. I come in to understand what you’re actually running, what’s getting in the way, and what the real goal is. The categories below are the menu of what I might recommend after that — not the menu of what I sell.

// 01 · Integration

Make the platforms you already pay for talk properly.

Most stacks already include the right pieces. The job goes wrong at the seams. ServiceM8 syncs invoices to Xero, but the tax codes don’t quite map. AroFlo sends totals across, but retention sits in a separate workflow. Buildxact pushes estimates, but progress claims need re-keying.

The default integration covers the easy 80%. The remaining 20% is where margin leaks and where the office spends its time. I write the bridge code for the specifics — your tax setup, your job structure, your edge cases — and tie it back to the platform APIs you already pay for.

The kind of work this looks like

  • ServiceM8 ↔ Xero quote and invoice sync with correct tax handling for your account setup.
  • AroFlo ↔ Xero retention and progress-claim mapping that survives a real construction job.
  • Buildxact estimate-to-Xero invoice handoff that holds up under variation orders.
  • Custom API wiring between your CRM, field tools and accounting.

What it’s not

I don’t do generic ‘set up your accounting integration for you’ work — there are bookkeepers and platform-certified setup specialists for that. Integration here means the custom layer that sits between platforms when the off-the-shelf connector doesn’t fit.

// 02 · AI workflow automation

Specific agents for specific named tasks.

AI is useful when it’s pointed at a specific task. Not ‘AI for your business’ — an agent that reads a particular kind of inbound email and drafts a particular kind of output, then puts it in front of a human for one click of approval.

The work is in the framing: what’s the named task, what’s the tolerance for error, where does the human stay in the loop, and what happens when the agent gets it wrong. I build agents that fit how your office actually runs, not the demo on the platform vendor’s website.

The kind of work this looks like

  • An AI agent that reads inbound enquiry emails and drafts a ServiceM8 quote for the office manager to review and send.
  • Document classification and routing for incoming supplier invoices — recognised, coded, queued for ApprovalMax.
  • Inbox triage that turns long email threads into structured job notes inside AroFlo or simPRO.
  • A first-pass review for variation requests, with the bookkeeper confirming before anything is posted.

What it’s not

I don’t sell ‘AI strategy.’ If a workflow doesn’t have a clear named task, an obvious cost saving, and a place for a human to approve the output, I won’t build it. Open-ended AI experiments are a fast way to spend money and get nothing back.

// 03 · Custom software

When off-the-shelf can’t deliver.

Sometimes the answer is build something. When your data lives in three places, when the report you need doesn’t exist in any of them, when the workflow your team actually runs is different from what every platform assumes — that’s when custom is right.

I build small, focused tools. Hosted dashboards, internal apps, portals between platforms. Things that own a single job and do it well, instead of replacing systems that already work.

The kind of work this looks like

  • A hosted dashboard joining AroFlo or simPRO with Xero for weekly job-by-job profitability.
  • An internal tool that tracks WIP positions properly and exports clean to Xero at month-end.
  • A subcontractor portal where subbies submit timesheets and progress photos against a job.
  • A client portal showing job status, invoices and progress claims in one view.

What it’s not

I don’t build replacements for ServiceM8, AroFlo, simPRO or Buildxact. Those platforms exist for a reason and they do their job well. Custom software here means the layer above or beside them, not instead of them.

// 04 · Strategy & discovery

Scope what’s worth building.

Sometimes the right first step is a paid scoping engagement. Two or three weeks of conversation, document review, and stack inspection, ending with a one-page recommendation you can act on — even if you don’t end up working with me to build it.

Discovery is the same shape whether or not it leads to a build. I want you walking away with a clearer picture of what’s worth fixing, what isn’t, and what the smallest version of a fix looks like. If the answer is ‘buy this off-the-shelf,’ I’ll say so.

The kind of work this looks like

  • A paid stack audit that identifies where the data is leaking margin.
  • A ‘build vs buy’ review for businesses considering a platform change.
  • A one-page recommendation: what to fix, what to leave, in what order.
  • A discovery sprint scoping a single integration or AI workflow before committing to build.

What it’s not

Discovery isn’t a free sales call. The output is a written recommendation you own — including the parts where I tell you not to spend money with me. If you’re looking for a free pitch deck, this isn’t that.

// How engagement works

Paid discovery. Fixed-scope build.

Discovery is paid. Build work is fixed scope, fixed price — agreed before any code is written, with payment milestones tied to deliverables. No retainers, no hourly billing on the public page. If a retainer makes sense for ongoing work after a build ships, we’ll talk about it then.

// Investment ranges

// Integration

Single-system

Low five figures.

// Workflow build

Multi-system or AI

Mid five figures and up, depending on scope.

// Discovery

Strategy engagement

Scoped per project, typically a few thousand dollars.

Final scope and price after discovery. No surprises.

// Next step

Tell me what’s broken. I’ll tell you if I’m the right fit.

Start with a discovery call