How an Engagement Actually Runs
Three illustrative scenarios showing what we look at, how we decide what is worth doing, and how success gets measured — drawn from the kinds of Australian businesses we work with.
About these scenarios
These are illustrative composites, written to show how we approach an engagement. They are not real named clients, and nothing on this page is presented as an achieved result or a guaranteed outcome. Our client work is commercially confidential and we do not publish client names, findings or numbers without written consent. If you would like to talk to a reference, ask us on the call and we will arrange it directly where the client has agreed.
A 40-staff building services firm
The Situation
The owner is fielding a steady stream of enquiry calls that reception cannot always reach, quotes are being written by hand in the evenings, and job data lives across a job management system, a shared drive and several spreadsheets. Leadership has been pitched an AI product by two vendors and has no way to judge either. Their question is not "what can AI do" — it is "which of these five problems is actually worth spending money on first?"
How We Would Approach It
- Map the enquiry-to-invoice process end to end and time each step
- Quantify what is being lost to missed calls and slow quote turnaround
- Review data quality in the job management system before assuming any of it is usable
- Score every candidate use case against value, effort and risk
- Sequence the roadmap so the first project is visible, low-risk and self-funding
What an Audit Typically Concludes
In a business shaped like this, the highest-value opportunity is usually not the most exciting one. Answering and triaging inbound enquiries reliably tends to beat anything involving predictive analytics, because the process is well understood, the data is simple and the outcome is measurable within weeks. The roadmap generally recommends deferring anything that depends on the messy spreadsheet data until the underlying process is cleaned up.
How Success Would Be Measured
Every recommendation carries its own measure, baselined before anything is built. These are the metrics — not results.
- Percentage of inbound enquiries answered and captured
- Median time from enquiry to quote sent
- Hours per week of after-hours administration
- Quote follow-up completion rate
A 120-staff wholesale distributor
The Situation
The finance team is manually re-keying supplier invoices into the accounting system, the operations team is copying order data between a portal and a spreadsheet, and staff have quietly started pasting customer information into public chatbots to draft emails. The board has asked management for an AI position paper. Nobody in the business has a mandate to write it.
How We Would Approach It
- Workshop with finance, operations and customer service separately — the shadow AI story only surfaces without leadership in the room
- Inventory every repetitive data-movement task and its true cost in hours
- Assess integration options against the accounting and ERP systems already in place
- Draft an AI use policy and data-handling rules sized for an Australian mid-market business
- Produce a costed roadmap and a board-ready position paper in plain English
What an Audit Typically Concludes
Businesses at this size almost always have two distinct problems running in parallel: a genuine automation opportunity in the back office, and an unmanaged governance exposure from staff already using AI informally. The governance work is usually cheaper and more urgent than the build. It is also common for the best answer to one or two of the "AI" candidates to be plain integration with no model involved — we say so rather than dressing it up.
How Success Would Be Measured
Every recommendation carries its own measure, baselined before anything is built. These are the metrics — not results.
- Invoice and order lines processed without manual re-keying
- Error and exception rate versus the manual baseline
- Staff hours returned to higher-value work each week
- Proportion of AI tool use covered by an approved policy
A 25-staff professional services practice
The Situation
The practice ran an AI pilot last year. It demonstrated well, everyone was impressed, and it never made it into daily use — the tool sat outside the workflow, so using it was extra work rather than less. Leadership is now sceptical and needs to know whether the earlier attempt failed because of AI or because of how it was introduced.
How We Would Approach It
- Post-mortem the previous pilot honestly, including the adoption and ownership questions nobody asked the first time
- Identify the specific tasks where AI removes work rather than adding a review queue
- Test the shortlist against the practice’s real documents and real constraints, not a demo dataset
- Build the highest-value item directly into the existing workflow, not alongside it
- Run role-specific training and appoint an internal owner before go-live
What an Audit Typically Concludes
Failed pilots are rarely a technology failure. The usual cause is that the tool created a new queue for a human to check, so it never saved anyone anything and adoption quietly died. The fix is to narrow the scope until the agent genuinely completes a task end to end, then wire it into the software the team already opens every morning. A smaller thing that completes beats a larger thing that needs supervising.
How Success Would Be Measured
Every recommendation carries its own measure, baselined before anything is built. These are the metrics — not results.
- Weekly active use by the team the tool was built for
- Tasks completed end to end without human rework
- Turnaround time on the target task versus the manual baseline
- Adoption at 30, 60 and 90 days post go-live
Why There Are No Logos on This Page
Consulting work touches commercially sensitive material — how a business quotes, where it is inefficient, what its data actually looks like. Most clients would prefer that stayed between us, and we think publishing a wall of logos to win the next deal is a poor reason to test that.
So we do it the other way around: everything on this page is clearly marked as illustrative, and if you want proof, ask for a reference on the call. Where a client has agreed to speak to prospective clients, we will connect you directly — which is worth considerably more than a testimonial we wrote ourselves.
What Would Your Scenario Look Like?
Book a free consultation and we will walk through your situation the same way — honestly, and with no obligation to go further.