How can you identify processes worth automating?

Start with the work, not the tool

Start with a familiar task and follow it from its actual beginning. A request may arrive by email or through a website form, be copied into an Excel or Google Sheets spreadsheet, and later be entered into another program or document. Note who handles each step, where information is entered again and what happens at the end. This reveals the real working process, including handoffs that may not appear in a formal procedure.

Look for repeatable patterns

A process may be a useful automation candidate when its inputs are familiar, its steps happen repeatedly and its outcome follows clear rules. Repetition alone is not enough: first check whether the work is already understood and whether avoidable steps can be removed.

• Details move from an email or website form into a spreadsheet, then get entered again somewhere else.
• A person checks a predictable condition and then performs the same next step.
• Updates or reminders are routinely sent after a known event.
• A recurring report is assembled from the same sources and follows the same structure.

Classify the work before choosing a solution

Repetition is only one signal. Separate frequent work with stable rules from tasks that produce many exceptions or depend on experience. A task may be repetitive but still need a person to judge unusual cases. Automate predictable preparation and routing; keep decisions with a person when context or accountability matters.

• Repetitive work: the same steps recur, but first check whether they can be simplified.
• Rule-based work: the same conditions lead to the same outcome, so a fixed rule may be enough.
• Exception-heavy work: missing or unusual information is common; improve routing and exception handling before automating the main path.
• Work that needs human judgement: decisions depend on experience or accountability; prepare the information but leave the decision to a person.

Ask a few practical questions

For one recurring task, note how often it happens, what rules guide it, what exceptions occur, where its information comes from and who makes the decisions. A useful first candidate happens often, has a clear result and manageable exceptions. If the rules are unclear or judgement dominates, improve those parts first.

Check exceptions before choosing a solution

Ask what happens when information is missing, a request is unusual or a system is unavailable. Decide who should see an exception and how the process can resume. A clear fallback is part of the workflow, not an afterthought.

Choose a useful first step

For example, a small business might receive a service request by email, copy its details into a spreadsheet and later re-enter them in its accounting or service system. First identify which fields and checks are consistent, which requests are unusual and who resolves them. Automate only the stable handoffs, then review the exceptions with the people doing the work. Connect another system only when the process needs its information.