Pick a job of work
Not “operations.” A concrete journey: a lead becoming cash, a ticket becoming a resolved job, an invoice becoming paid, a hire becoming productive. One path, walked end to end.
Practice / Approach
The method
Follow the work, not the org chart.
Processes on a slide are almost never the processes in the wild. We trace a job of work from the moment it starts to the moment someone can honestly call it done.
Sit with the people who do it
Short interviews beat workshops. We want the unofficial route: the spreadsheet beside the CRM, the Slack DM that actually starts the job, the person who re-keys the same data because the tools don’t talk.
Find the drag
Where time pools. Where errors repeat. Where only one person knows the trick. Where a tool was bought to hide a broken step. The map names these in English, not in framework jargon.
Rank the path
Every recommended move gets a reason, a cost shape, and a way we’ll know it worked. Automate, replace, stop, or leave human. Sequence matters: we don’t pave the swamp before we drain it.
Build in slices
If we implement, we ship a working slice against the real work — not a sandbox demo. Exceptions are designed for, not discovered in week six. You see it running in your environment.
Hand off the trail
Credentials, code, the map, and a short runbook. The system should survive a vacation. If it only works while we’re in the Slack channel, we didn’t finish.
Principles
Strategy before software. Path before platform.
We would rather kill a step than automate it. We would rather use the tools you already pay for than introduce a fourth source of truth. We would rather write a boring rule than a clever model.
The output is a system your team can run on a Tuesday when nobody is watching — not a workshop high and a folder of Miro boards.
Next step
Bring us one stuck path.
You don’t need a full operations audit. Name the job of work that keeps landing back on your desk.