Good Automation Starts With a Boring, Repetitive Task

Automation sounds exciting when you talk about what it could eventually do.

You can imagine workflows connecting everything, responding to every event, and taking care of work in the background. Before long, the conversation includes branching logic, conditions, and a visual builder.

But the most useful place to start is usually much less interesting.

Someone opens a report every week, finds the same kind of people, and makes the same update.

That is a good automation problem.

It has a clear starting point, a repeated decision, and an action someone already understands. You can watch the manual process and ask what it would take to stop making that person do it.

Start with the work people already do

While shaping workflow automation at BrandChamp, the core idea was fairly straightforward: help admins make routine updates to ambassadors based on defined criteria.

An admin might want to find people who have not completed an activity recently and add an “Inactive” tag. Another might want to identify people who have completed enough activities to qualify for a particular group.

Without automation, someone has to remember to check, apply the filters, and update the right people.

Each individual task might be manageable. The effort comes from remembering and repeating it across an ongoing program.

That is where automation can earn its place.

You have a process that already exists, and you can describe what a successful run should accomplish.

A small first version can do useful work

The first version of our workflow automation focused on scheduled tag updates. Admins could define an audience, choose a schedule, and add, remove, or replace tags.

There were plenty of other actions we could consider later.

Keeping the initial scope focused gave us a concrete problem to solve. Could an admin set up a routine update, understand who it would affect, and trust it to run?

Even that requires care.

What does replacing tags mean? What happens when someone no longer matches the filter? Can the admin understand the effect of several actions in the same workflow?

A small action list does not automatically make the experience simple. You still have to explain what will happen.

The order of the setup matters

A system might think about automation as a schedule followed by a query and an action.

A person may think, “I want to do something with these ambassadors.”

That is why the audience deserves attention early in the setup. Once someone understands who they are targeting, choosing the action and timing becomes easier to reason about.

The words in the interface matter here too. A filter description needs to help someone predict which people will be included.

If an admin cannot confidently explain the audience they have selected, asking them to automate an update to that audience is a big step.

People need to see what happened

Automation removes the need to do the work manually. It also removes the direct visibility that comes from doing it yourself.

That makes logs part of the experience.

An admin needs to know whether a workflow ran, what it did, and whether anything failed. If several actions were configured, the results should help explain all of them.

Otherwise, the next task becomes checking whether the automation worked.

There are also ordinary calendar details to resolve. A monthly schedule set for the 31st needs predictable behaviour in shorter months. These decisions are easy to miss in a mockup, but they affect whether the product feels dependable.

I like automation problems that can be described in a plain sentence.

“Every week, find these people and make this update.”

Once that works well, there is room to do more. The first milestone is giving someone one less routine task to remember.