How to Map a Business Workflow

Before anyone automates anything, someone has to write down what actually happens today — not what the process is supposed to be, what it actually is, including the workarounds. This is the method, on its own, with no tools required to start.

1. Pick one process, not a whole department

Workflow mapping fails when it's too broad. "Sales" isn't a workflow. "What happens between a website inquiry and the first reply" is. Pick something with a clear trigger (an event that starts it) and a clear end state (when it's considered done).

2. Write down every step, in order, including the boring ones

Not "we follow up with the lead" — the actual sequence:

• The inquiry arrives in the inbox
• Someone notices it (when, exactly — immediately, or whenever they next check?)
• They copy the details into the CRM
• They decide whether it's worth pursuing
• They write and send a first reply
• They set a reminder to follow up, or don't

The steps people skip when describing a process out loud — "and then I just check back later" — are usually where the actual time and inconsistency live. Write those down too.

3. Note who or what performs each step, and how long it takes

For each step: is it a specific person, whoever's free, or a tool? How long does it actually take — not the best case, the typical case, including the time spent waiting for someone to get to it? This is what turns a vague sense of "this takes too long" into a specific, defensible number.

4. Mark the decision points

Somewhere in most workflows, a human makes a judgment call: is this lead worth pursuing, is this application worth a screen, does this client need a call instead of an email. Mark these separately from the mechanical steps around them — they're the parts that usually shouldn't be automated away, even if everything around them can be.

5. Sort each step into "automate" or "keep human"

A simple, reliable filter: steps that are repetitive, predictable, high-volume, rule-based, or purely administrative are strong automation candidates. Steps that are strategic, sensitive, creative, relationship-driven, or a high-value decision usually aren't — not because AI can't touch them, but because the judgment involved is exactly what you're paying a person for.

6. Prioritize by volume × time, not by how annoying it feels

A five-minute task done twice a week matters less than a two-minute task done fifty times a week. Multiply frequency by time spent, and you'll often find the highest-value automation candidate isn't the step that feels most tedious — it's the one that quietly repeats the most.