Twenty-Eight Questions Before You Sign

Usually yes, and being asked to is a good sign rather than a bad one. A paid discovery that produces a documented process map, a list of exceptions and a firm build quote is far cheaper than a fixed-price build scoped from a one-hour conversation, because the latter either carries a large risk premium or turns into variations later. What matters is that discovery produces something you own and could hand to a different vendor: if the output is a proposal rather than documentation, it was a sales exercise you paid for.

It depends on volume and complexity rather than on preference. Established platforms are quicker to build on, easier for your own team to maintain, and cheaper at low to moderate volume, but per-task pricing becomes expensive at scale and complex logic gets unwieldy. Custom builds cost more upfront and need someone to maintain them, but they handle complexity better and cost less per transaction at volume. Ask any vendor to justify their recommendation in terms of your volumes specifically, and be sceptical if they only ever recommend the approach they happen to specialise in.

For a single well-scoped process in a business with reasonable documentation, two to six weeks from signature to live is a realistic range. Anything quoted at under a week is either genuinely trivial or has not been scoped properly. Anything running beyond three months for one process usually indicates scope that expanded during the build, which is a sign discovery was inadequate. Multi-process programmes should be sequenced as a series of short projects rather than a single long one, so value arrives early and lessons carry forward.

The Australian market varies enormously, from a few thousand dollars for a straightforward integration on an existing platform to well into five figures for a complex multi-system workflow with custom logic. Rather than benchmarking against a headline figure, get every vendor to quote the same clearly documented process and compare the totals including ongoing costs. A quote that is dramatically lower than the others usually reflects a narrower reading of the scope, so check what each has assumed about exception handling before treating them as comparable.

For straightforward integrations on established platforms, a capable internal person can absolutely build and maintain them, and doing so gives you the fastest change cycle and the lowest ongoing cost. The realistic constraints are capacity and continuity: internal builds tend to stall when the person who built them gets busy, and they become a liability if that person leaves without documentation. If you go internal, insist on the same standards you would demand of a vendor, documentation, monitoring, and a service account rather than a personal login.

This is exactly why the ownership questions matter more than any other section of this list. If the workflows run in your accounts, on your licences, with documentation you hold, a vendor disappearing is an inconvenience, another provider can pick them up. If everything lives in the vendor’s environment, their failure is your outage, with no notice and no path to recovery. Ask the question directly, get the answer in writing, and treat a vague response as a material risk rather than a formality.