Skip to content
SwitchyardAI automation agency · Australia Book 15 min

Costs

What an AI automation engagement costs

Published prices for each stage, what moves a project up or down the range, and the running costs to plan for.

The figures

Engagement prices by stage
StageLengthPrice (AUD)Deliverable
AI opportunity auditAbout 2 weeksFrom $2,500; typically $2,500–$5,000Process map, opportunities with ROI estimates, risk assessment, 90-day plan
Pilot30–60 daysFrom $5,000; typically $5,000–$15,0001–2 working automations, training, dashboard, measured ROI report
Scale and optimiseOngoingCustomExpanded coverage, monthly reviews, support and training

These are SG1 Consulting’s published figures for its audit, pilot and scale stages.

What moves an audit up the range

  • Number of teams involved. One team with one process is quicker to map than three teams that hand work to each other.
  • Undocumented processes. Common and fine, but mapping work that lives in people’s heads takes more interviews.
  • Number of systems. Each system that holds part of the process needs to be understood, including what its integrations allow.
  • Compliance requirements. Health, financial and legal work needs a more careful look at data flows and approvals.

What moves a pilot up the range

  • Integration difficulty. Systems with good, documented connections are faster to work with than older software that needs workarounds.
  • Document variety. Reading one supplier’s invoice layout is simpler than reading forty.
  • Exception handling. The more ways the process can go sideways, the more design and testing it needs.
  • One automation or two. A second automation that shares systems with the first costs less than one that starts fresh.

Running costs to plan for

Beyond the engagement itself, automations have running costs that you pay to the providers directly. The audit estimates them, but the categories are worth knowing now:

  • Automation platform subscriptions, usually priced by the number of runs, tasks or operations per month.
  • AI model usage, usually charged by the volume of text processed. For typical business admin volumes it is often modest, but large documents or high volumes add up.
  • Extra licences in systems you already own, if an integration needs a higher plan.
  • Maintenance time, whether your own staff’s or an ongoing arrangement under the scale stage.

We don’t quote other companies’ prices here because they change often; the audit lists current figures from each provider’s own pricing page for the tools that suit you.

Working out whether it pays

The honest way to judge an automation is the same arithmetic you would use for a hire. Take the hours the process consumes each week, the cost of those hours, and the cost of the delays and errors it causes. Compare that with the engagement cost plus a year of running costs. If the process happens rarely, or the time it takes is small, the numbers usually do not work, and the audit will say so.

Illustrative example, not a client

A business re-keys around 60 supplier invoices a week from email into its accounting software, at roughly four minutes each including checking. That is about four hours a week. An automation that extracts the details and queues each invoice for one-click approval might bring the human part down to under a minute per invoice. Whether that justifies a pilot depends on the hourly cost, the error rate today, and the running costs, which is exactly what the audit works out.