What Automation Actually Does Well - Well for me at least

When I pulled AI back from my writing, I didn’t stop using it. I just moved it to a different part of my work. That shift led me much more seriously into automation, and it’s an area I’ve come to really enjoy, even when it frustrates me.

I’ve been using Zapier for a while now, mostly by experimenting and learning as I go. Over the last six to twelve months it has changed rapidly, adding new capabilities at a pace that can be hard to keep up with. There are moments where it feels like you finally understand how something works, only to realise the platform has already moved on again.

Instead of seeing that as something to fear, I’ve made a decision to lean into it. I genuinely believe that people who become good at automation, not just using it but understanding it, will be able to work meaningfully for a long time. Not because automation replaces people, but because it reshapes where human effort is most valuable.

What surprised me early on is that automation is not a purely technical exercise. It requires real creative thinking. Tools like Zapier offer templates and pre-built automations, and those are useful starting points, but the real value comes when you start automating your own work. That means paying attention to what you actually do, where the friction is, and what decisions repeat often enough to be supported by a system.

That process takes time. You have to think carefully about triggers, actions, and outcomes. You have to test and break things. You have to refine prompts and logic. If you’re using AI-based steps, you also have to be mindful of cost. Automations that rely on models from OpenAI and similar providers often use tokens, and if you’re not paying attention, expenses can creep up quietly.

None of that is a reason not to automate, but it is a reason to do it deliberately.

A good example of where this has worked well for me is client intake. I started with a simple goal. Whenever a new client reached out, whether that inquiry turned into a contract or not, I wanted their information stored in one place. I use Notion for this, and I set up an automation so that entering basic client details would trigger the creation of a client file.

From there, that file becomes the foundation for everything else. Documents, quotes, notes, timelines. The administrative load drops significantly, not because the work disappears, but because the setup work does. I’m no longer rebuilding the same structure every time.

I’ve also been using automation to help unpack large amounts of data I collect through my work. The system can summarise, organise, and map information against deliverables and outcomes. That doesn’t mean I take those outputs at face value. It’s much more like being handed a report by a staff member. I still need to read it, question it, and decide what to do with it. Automation doesn’t remove responsibility. It just changes where my energy goes.

This is where I think automation really shines. It reduces load without touching judgement. It handles repeatable tasks, preparation work, and administrative complexity, leaving me with more capacity to think, decide, and write.

It’s also where I see a clear difference between automation and AI-generated content. When AI writes for you, there’s a real risk that your thinking gets thinner. When automation supports you, your thinking often gets stronger because you’re less distracted by everything around it.

That doesn’t mean automation is easy or always satisfying. It can be fiddly and frustrating. Things break. Logic loops don’t behave the way you expect. Costs need monitoring. But the effort is front-loaded. Once something works, it keeps working, and that reliability matters more than speed.

I don’t see automation as a way of doing less thinking. I see it as a way of protecting the thinking that matters. Writing, judgement, and sense-making are things I want to stay close to. Administration, coordination, and repetitive setup are things I’m very happy to design once and hand over.

That balance has made my work feel more sustainable. The tools are still there, and they’re still powerful, but they’re no longer sitting in the wrong place in the system.

This is where I want to keep going next. Not chasing tools for their own sake, but continuing to work out how automation can support good work without eroding responsibility. That feels like the real opportunity, and the real discipline, in this phase of AI.