Article
If you automate a really dumb step in a workflow, you usually end up with a faster dumb step. The core problem is still sitting there, only now it has more machinery around it.
I ran into this while building Sales Scout, a tool that helps people get ten prospects a day. It is supposed to be an action tool. You use it to build momentum and get to the work that matters.
At one point, I put a chat agent inside the product. I thought I was being pretty clever. You could talk to it, configure parts of the tool through it, and the agent could even raise its hand when it needed help and ask you to come online.
It sounded useful. In practice, every extra interaction added complexity.
Instead of helping people reach the end result, the chat agent created more steps between them and that result. I was struggling with the same problem while building the product. There was more to configure, more to respond to, and more interaction that did not need to be there.
So I cut the agent.
The results have been way, way better.
Start with the step, not the automation
When a workflow feels slow or awkward, automation can look like the obvious fix. But the first question is whether the troublesome step belongs in the workflow at all.
Look at what that step contributes to the end result. If it adds interaction without making the outcome better, automating it will not clean up the workflow. It will preserve the same complexity and help it move faster.
That is what happened in Sales Scout. The product's job was clear: help someone get ten prospects a day and build momentum. The chat agent did more things, but those things made it harder to get to the result. More capability did not make the product better.
Removing the feature did.
Deletion is a real workflow decision
Getting rid of a step can feel less sophisticated than automating it. There is no agent to configure and no extra interaction to manage. You are left with a shorter workflow that asks less of the person using it.
That can be exactly what the work needs.
Before you automate a clumsy step, pressure test the step itself:
- Does it move the person closer to the intended result?
- Does it improve the product, or does it create another interaction?
- Would the workflow work better if the step disappeared?
If the step does not need to be there, do not automate it and do not overthink it. Get rid of it. That is usually the cheaper option anyway.