The workflow audit that finds the tasks nobody realised were automatable

Ask any owner "what would you automate?" and they'll name something big and obvious. Invoicing. Payroll. Customer service.

But the workflows that most reward automation are almost never the ones people name. They're the boring, invisible ones. The one your admin does every Thursday afternoon for two hours. The report your operations manager pieces together every month from four systems. The onboarding checklist that half the time gets missed halfway through.

The workflow audit is the entry point to our AI & Automation service — literally sitting with the people doing the work and mapping where automation earns its keep.

What the audit actually looks like

• 90 minutes with the person doing the work, not the person managing them.
• We watch the actual task. Not a description of it. Not the ideal version. What actually happens on a Thursday afternoon.
• We map the steps. Every click. Every copy-paste. Every "then I look at this spreadsheet and cross-reference."
• We score each step. Automatable now. Automatable with tooling. Needs a rethink first. Genuinely human.
• We rank the opportunities by hours saved, error risk reduced, and effort to implement.

Why we insist on "watching the actual task."

Because the description you'd give in a meeting is almost never how the work actually happens. There are workarounds nobody remembers doing. There are three-way handoffs between systems that were meant to be temporary in 2022. There are Post-it notes on monitors. There are entire steps that exist because "the report needed it that way" for a client who left last year.

You can't automate what you can't see. And you can't see it from a management deck.

The tools we use for the "automatable now" bucket

• Power Automate for anything sitting inside the Microsoft 365 world.
• Zapier-style connectors for anything that reaches across SaaS tools.
• Custom AI-assisted workflows where the task involves interpreting text or extracting from documents.

The tools we sometimes recommend against

• Anything that requires the source data to be perfectly clean when it isn't. Fix the data first, automate second.
• Anything that would remove a human checkpoint from a client-facing process.
• Anything where the setup cost outweighs the annual saving. We'll say so. Not everything's worth automating.