Spotting Automatable Tasks at Work

TL;DR - Most people don't know where to look for automation wins. A quick task audit finds them: score each recurring task, then write a one-line spec for the best candidates.

Why it matters

You can't automate what you haven't spotted. A simple audit of your week turns "AI sounds useful" into a concrete list of things to build.

Worked example - the audit

For each recurring task, score 1-4:
              Frequent? Same steps? Clear trigger? Low judgment?
Copy leads to CRM   Y        Y            Y             Y   -> automate
Write weekly update Y        Y            Y             N   -> AI-assist
Approve refunds     Y        N            Y             N   -> keep human

Steal this - the automation spec

For each candidate, write one line:
Trigger:  [what starts it]
Steps:    [what happens, incl. any AI step]
Output:   [the result]
Risk:     [what could go wrong + the guardrail]

Common mistakes (and the fix)

• Looking for big, glamorous automations. Fix: the boring repetitive stuff has the best ROI.
• No spec before building. Fix: the one-liner makes the build obvious.
• Ignoring risk. Fix: name the failure mode up front.

Good to know

This audit is exactly how consultants find automation ROI in a business. Do it on your own week first; the same method scales to a team or department. The specs you write here become the builds in the next lessons.