Repeatable Workflows

Spotting what repeats

List the tasks you or your team redo every week: reporting, email triage, follow-ups, summaries. These are your best candidates. Start with the most painful one, not the most complex. Painfulness is an excellent indicator: a hated task is a task you postpone, rush, or do in a bad mood — automating it pays back time and morale.

For Marc, the inventory took twenty minutes in a team meeting: everyone listed what they redo each week and rated it 1 to 5 on annoyance. The winners: the monthly follow-up with clients late on documents, the weekly check-in minutes, and the vendor invoice reconciliation. Three processes, all perfect for starting — frequent, well-bounded, with no binding decision in the middle.

You'll notice that everything you've learned in the previous chapters converges here. A repeatable workflow is simply a good brief (chapter 1), possibly fed by monitoring (chapter 2), producing documents (chapter 3) or analyses (chapter 4) — but written once to be replayed every week. This chapter turns your one-off successes into a system.

Not all processes are equal

This sorting saves you the classic disappointment: trying to automate the firm's most complex process right away, failing, and concluding that "it doesn't work for us". The winning progression is the reverse: a first simple workflow that runs reliably builds confidence and appetite — in you as in your team. The complex processes will come, carried by that momentum.

Documenting the process once

Describe your process step by step and ask Claude to turn it into a reusable checklist. Next time, you paste the checklist and it executes it, without re-explaining everything. It's the same principle as the written procedures of a well-organized firm — except the procedure isn't just readable: it's executable.

Describe the process the way you'd explain it to someone covering for you during your vacation: where the information arrives, what you do with it, in what order, what the final result looks like, and the special cases ("if the client has already been chased twice, the tone changes", "clients in legal proceedings are never chased"). These special cases are the salt of the workflow: they're what makes the difference between a naive automation and an automation that sounds like you.

Here is how I prepare the monthly follow-up for clients late on documents:
1. I export the list of incomplete files from our software
2. I set aside clients in legal proceedings or disputes
3. For a 1st reminder: courteous email with the list of missing documents
4. For a 2nd reminder: firmer tone + deadline + partner CC'd
5. I keep a tracking table of reminders sent
Turn this into a reusable checklist I can give back to you each month with my export, so you execute everything: sorting, personalized emails per client, and an updated tracking table.

flowchart LR
  D["Document the process"] --> CL["Reusable checklist"]
  CL --> E["Execution by Claude"]
  E --> M["Measure the time saved"]
  M -->|"Improve"| D

Test, stabilize, improve

A workflow isn't born perfect: it becomes perfect. For the first two or three runs, compare the result with what you would have done yourself and note the gaps: a tone too curt here, a forgotten special case there. Then feed those corrections back into the checklist itself — not just into that day's conversation. That's the difference between correcting one essay and improving the textbook.

The diagram above isn't decorative: the "Improve" arrow is the step everyone skips, and yet it's what makes everything work. A checklist enriched with three months of real cases becomes remarkably reliable — it now contains more business knowledge than the head of any new hire. At Marc's firm, the follow-up checklist is on version 4: it handles nine special cases that no initial brief could have foreseen.

Build your workflow library

Since chapter 1, you've been setting aside your best briefs. It's time to turn them into a team tool: a single shared document — the firm's workflow library — where each entry fits in four lines: workflow name, when to use it, the checklist to paste, and the last update date. Nothing more: sophistication would kill adoption.

The effect of this library goes beyond individual time savings: it levels up the whole team. The junior colleague's client follow-up comes out as well calibrated as Marc's, because the same checklist is running. And when someone improves a workflow, the whole team benefits on the next run. It's the most durable asset this course has you build: it appreciates over time instead of wearing out.

Bringing your team on board

The limiting factor is almost never the tool: it's adoption. Three practices make the difference. First, show rather than explain: a ten-minute demonstration on the task everyone hates beats a one-hour meeting about "AI at the firm". Second, let everyone choose their first workflow: adoption comes from a personal pain relieved, not from a directive. Finally, institute a light ritual: five minutes at the weekly check-in where everyone shares one success or one failure with the tool that week.

Also anticipate the fears, because they're legitimate: "will this replace me?" deserves a real answer. At Marc's firm, the answer is factual: nobody was replaced, but the evenings spent finishing follow-ups have disappeared, and the recovered time has gone toward client advisory — the part of the job everyone prefers and that bills better. An honest account of what the tool does (the mechanical tasks) and doesn't do (the relationship, the judgment, the signature) defuses most of it.

Measure to convince

Note how long the task took before, and after. For Marc, adding up those minutes across the team and across the year is what turns "AI gadget" into "15 hours recovered per week". The measurement doesn't need to be scientific: an honest before/after estimate per workflow, multiplied by the frequency and the number of people involved, is largely sufficient.

Run the numbers on the monthly follow-up case: 3 hours before, 30 minutes after (verification and sending included), that's 2.5 hours saved per month, 30 hours per year — for a single workflow, a single person. Marc's dashboard fits on one page: each workflow, its estimated monthly gain, its status (testing / reliable / needs improvement). Total after four months: 14 to 16 hours per week across the team of four. The course's goal wasn't a slogan.

And this measurement has one last merit: it tells you where to invest next. If the monthly monitoring only saves one hour while the reporting saves six, you know which of the two deserves a version 2 of its checklist. Measuring is steering — exactly what Marc does for his clients, this time applied to his own time.