AI Discovery and Automation
Map your highest-effort manual processes into an automation-ready workflow — in one 60-90 minute session.
Not "AI implementation." This is the step before that — finding out exactly what's worth automating, and what isn't, before anyone builds anything.
Who this is for — operations, admin, finance and logistics teams doing the same manual work every week (reconciliation, invoice matching, quoting, approvals, onboarding) where nobody's ever actually mapped out what the process really is.
Get the $49 Field GuideThe problem
Most manual processes were never designed. They accumulated — one exception handled a certain way becomes the rule, one spreadsheet becomes the system of record, one person becomes the only one who knows how it actually works.
The cost isn't always obvious in hours. It shows up as: work that only happens because one person remembers all the steps; exceptions handled differently depending on who's covering that day; reporting that's always a day behind because someone has to compile it by hand; nobody being able to say, precisely, what "normal" looks like.
You don't fix any of that by buying AI tools. You fix it by mapping the process first — properly, including the awkward cases — so you know what you're actually looking at.
Real example: mapping a manual process
The problem
A Sydney logistics client reconciled roughly 2,000 containers a week against invoices from about 30 contractors, entirely by hand.
The session
One real reconciliation observed start to finish, plus the exception case: an invoice approved from memory, with no ticket or signed proof of work.
The map
What triggers the process, which systems it touches, where the real judgement calls are, and what "done" actually means — all written down.
The result
A process map that had never existed before. No automation has been built yet — that's a separate, later step.
Real example: accelerating a sales process
The meeting
A first-call meeting with a new prospect, captured and transcribed.
The brief
A tailored sales brief drafted the same day, built directly from specifics raised in the actual conversation, not a generic template.
The proposal
A 16-slide proposal deck, a matching proposal document, and a scope schedule, all built from one shared content source so everything stays consistent.
The result
A complete, tailored proposal pack, built directly from what was actually said in the meeting.
The method — "AI Discovery and Automation," in seven steps
- 1Pick a process with real volume and real pain — not a hypothetical one.
- 2Define the unit of work — what counts as "one" of these, and what "done" actually means.
- 3Map the current flow, including the exceptions. The exceptions are usually where the real cost lives.
- 4Identify the data sources and system boundaries — what touches what.
- 5Define the controls — approvals, audit trail, who's allowed to decide what.
- 6Design the future-state workflow — what it should look like once it's mapped.
- 7Score it for automation readiness — which parts are fixed rules, which are real human judgement that should stay human.
What's in the guide
- The field guide — the exact document used to run this.
- A mapping template you fill in as you go.
- The conversation guide — how to run the session with someone doing the work, not just guess at it from your own assumptions.
- Example outputs from the logistics case above, so you can see what "done" looks like before you start.
Who shouldn't buy this
Honestly — this isn't for everyone, and it's better you know that now:
- —you're looking for a magic AI prompt that does the thinking for you (this is a method, not a shortcut);
- —you don't have a process that actually repeats (if every case is a one-off, there's nothing here to map);
- —you can't get time with the person who actually does the work (the exercise needs them, not just you describing it secondhand);
- —you want it built for you without anyone internally being involved (that's a different, bigger engagement, not this).
Frequently asked questions
Get the field guide
Get the $49 Field GuideIf you'd rather have this done for you once you've mapped it, we can help with that too — after you've run the discovery yourself.