A map of how the work actually runs
Every step, who touches it, what triggers it, what it waits on, and where it goes when something is unusual. Drawn from watching and interviewing the people doing the job, not from the procedure document.
Process Analysis & Discovery
Find out what is actually broken before you build anything
A paid discovery engagement that maps how work really happens, prices the problem, and tells you honestly whether building is worth it.
Overview
Most software that disappoints was built to the wrong brief. Not badly built — built precisely to a description of the work that nobody had checked. A business process analysis consultant in Australia is worth hiring for exactly one reason: to find out what is really happening before money is committed to changing it.
This is the first step in every Drawn AI engagement, and we charge for it. Free discovery is a sales meeting with a different name, and a sales meeting has an obvious incentive to conclude that you should buy something. We would rather be paid to give you an answer you can act on, including when that answer is that you should not build anything.
The gap we are looking for is between how work happens and how the manual says it happens. Almost every business has one, and it is not a sign of poor management. Processes drift. People invent sensible workarounds for systems that stopped fitting. Steps survive because a piece of software could not do something years ago, long after that limitation went away. None of it gets written down, and none of it appears in a requirements document — which is why so much custom software solves the documented problem rather than the real one.
Liam and Jacob founded Drawn AI in 2024 out of retail operations management and mining operations. That is the background this work draws on: standing in the operation and watching what people do, rather than reading what they were told to do.
Capabilities
What we build
Where the time goes
We measure elapsed time as well as effort. Most processes spend the overwhelming majority of their duration waiting — for an approval, a file, a callback, someone's return from leave. Teams routinely misjudge which step is the slow one, because the step that feels painful and the step that costs days are usually different.
Where errors get in
Every point where data is re-keyed, transcribed, interpreted or remembered. These are also the points where rework, credit notes, disputes and compliance findings originate, so they usually cost more than the time they consume.
A prioritised list of what is worth fixing
Ranked by frequency, cost of the current failure, and effort to change. Some items are configuration changes in software you already own. Some are policy decisions that cost nothing. Some are genuine build candidates. They are not presented as one undifferentiated wish list.
An honest recommendation, including "don't build this"
Sometimes the answer is a setting you already have, a process change, or a product you should just buy. We say so. A recommendation that never says no is not a recommendation.
A scope you can price, from us or anyone else
If building is justified, you leave with a specification concrete enough to take to any developer. It is yours. We would rather compete on the build than hold the map hostage.
Signs this is worth looking at
- Nobody in the business can describe the process the same way twice, and each version is partly right.
- Your documented procedure was written for a system you no longer use.
- You have been quoted for custom software but cannot tell whether the quote solves the actual problem.
- The same failure keeps recurring and each fix addresses the symptom rather than the cause.
- One long-serving person is the only reason the process works, and their retirement is a business risk.
- You have already bought software that was supposed to solve this and it is barely used.
- You can feel that a function is inefficient but cannot say where the hours are going.
How It Works
How we approach it
1
Discover
We agree what we are examining and why, then spend time with the people doing the work. Interviews, observation, and access to the systems involved — the accounting file, the job management system, the shared drive, the spreadsheets that quietly hold the process together.
2
Map
We document the current state, end to end, including the exceptions. The exception path is usually where the real cost lives, because it is where the experienced people spend their day.
3
Analyse and prioritise
We attach numbers to the map: how often each path runs, how long it waits, what the failures cost. Then we rank the opportunities against the effort to fix them.
4
Recommend
You get a written recommendation and a walkthrough. If it says build, it says what to build first and what to leave alone. If it says don't build, it says that plainly and explains what to do instead. Where discovery leads on to delivery, the usual next steps are workflow automation for process and approval problems, custom software development where a system genuinely needs to exist, or replacing spreadsheets with custom software where a workbook has outgrown itself. Our work in mining and professional services shows how this applies in operations where downtime and billable time carry very different costs.
Questions
Frequently asked questions
For a single process in a small or medium business, typically two to four weeks from kickoff to written recommendation. Broader reviews spanning several functions take longer. The variable is access — how quickly we can get time with the people doing the work and visibility of the relevant systems, not how fast we can write.
Discovery is priced and scoped separately as the first step. If you go on to build, that work runs on a monthly subscription with a three-month minimum term and 30 days' notice to cancel. If discovery concludes you should not build, the engagement ends there and you keep everything we produced. There is no obligation to continue.
A current-state process map, an analysis of where time and errors accumulate, a prioritised list of opportunities with the effort each requires, and a written recommendation. Where building is justified, that includes a scope specific enough to be quoted by any developer. All of it is yours to use however you like.
Yes, and it happens. Often the fix is a configuration change in software you already pay for, a change to who approves what, or removing a step that no longer serves a purpose. Being paid for the analysis rather than the build is what makes that recommendation possible to give.
Usually a few hours each from the people who do the work, spread across a couple of weeks, plus system access. We work around operational schedules — shift patterns, month-end, shutdowns — rather than expecting the business to pause for us. Most of the analysis happens on our side.
Not always, and if your requirements are genuinely settled we will tell you to skip it. But it is worth an hour's conversation first. The common pattern is that the requested build is real and the sequence is wrong, or that one assumption inside it does not survive contact with how the work runs.
Want to talk it through before committing to anything?
A first conversation costs nothing and usually ends with a clearer idea of what is worth building — sometimes that answer is “not yet”, and we will say so.
- No obligation and no sales sequence
- Built around your existing systems
- Australian-based, Australian-hosted data
- 3-month minimum, then 30 days’ notice
Start a conversation
No lock-in templates. Built around you.