Which manual work is worth automating.
A ranked list with payback against each item, the right platform named for each one, and a first build scoped tightly enough to fund from a line item.
3 weeks · fixed scope · a costed shortlist at the end
We count before we recommend.
Most automation money is wasted on the wrong tool for the process, or on a process that should have been fixed rather than accelerated.
- 01Volume: how often the work actually happens, taken from the system rather than from memory.
- 02Handling time, and how much of it is waiting rather than working.
- 03Variability: whether the input is the same every time or never twice.
- 04Whether the systems involved have an API, or whether a screen is the only way in.
- 05The exception rate, and what happens to an exception today.
- 06How many people touch it, and what it costs per transaction once you total that.
- 07Whether the process is stable, or about to change because the platform under it is being replaced.
- 08What breaks downstream when it goes wrong, which is usually the real cost.
What lands on your desk
A ranked opportunity list
Every candidate with volume, effort, payback and risk against it, ordered so the first one funds the argument for the second.
A platform recommendation per process
Power Platform, UiPath, integration or an agent — named for each one, with the reason and the running cost.
The ones to leave alone
Explicitly. Low volume, about to change, or better fixed than automated — with the reasoning, so nobody reopens it in six months.
A baseline
The numbers as they stand today, so whatever gets built can be measured rather than described.
A first build, scoped
One process specified tightly enough to quote, with the exception path and the measures agreed before anyone writes anything.
A capability view
What your own team could build and maintain, and what genuinely needs us — because the second list should get shorter over time.
Three weeks, one shortlist.
- Week one
Sessions with the people who do the work, and the system data behind what they describe. The two rarely match, and the gap is usually where the opportunity is.
- Week two
Sizing and platform fit for each candidate, including the running cost nobody counts and the maintenance profile of each option.
- Week three
Ranking, payback, and the scoped first build — walked through with the people who would have to live with it before it is finalised.
Before you ask
We already know which process we want automated.
Then we will confirm it quickly, or tell you why the one next to it pays back faster. Either answer is worth three weeks before a build.
Will you just recommend the platform you sell?
We build on all of them, so the recommendation has no stake in the answer — including when it is that a rule, a report change or a conversation with the vendor would do it.
Bring us the work that will not scale.
Forty-five minutes on what your team is doing by hand, and whether an assessment is worth running at all.