Not Everything Needs to Scale

One thing I've noticed since joining WriftAI is how often growth changes the answer to a problem that didn't seem like a problem before. By the time I joined, a lot of things had already changed.

If someone needed something, they usually knew who to ask. If something broke, someone fixed it. If a process felt unnecessary, we just didn't have one. There wasn't much reason to build a system around every little thing when the person you needed was usually sitting a few metres away.

That gets harder when the same few people aren't in every conversation anymore. More people are involved, more work is happening at the same time, and the shortcuts that once made everything easy can start creating their own problems.

Over time, WriftAI has added trackers, processes, automations, and a few more meetings along the way. And we've needed a lot of them.

But one thing that stood out to me pretty quickly was that some of the decisions people were happiest with weren't things they added. They were the things they looked at, considered building, and decided they could live without.

Growth creates a strange pressure to add things. More structure can feel like progress. More systems can make things feel more organised. Sometimes they are. Sometimes we're just making a simple thing more complicated.

Maybe We Didn't Need Another Form

One of the first things I noticed after taking over People & Ops was how simple some requests actually were.

Someone needed another screen for their desk, so they messaged me. I checked what we had, figured out what they needed, and sorted it out. Nothing particularly complicated.

Then I started thinking about how we were handling these things more generally.

Should there be a proper equipment request form? A place where people could submit what they needed, explain why they needed it, and maybe have the request go through an approval step before anything was ordered?

It would certainly look more organised. But then I thought about what the process was actually solving.

If someone needs a screen, I don't really need a workflow to tell me that they need a screen. I need to know what they need, whether we have one available, and get it to them.

So we left it alone.

There's no equipment request system. Someone asks, we figure it out, and the screen gets sorted.

It's a small example, but I've found myself coming back to it quite a bit. When you're trying to bring more structure to the way things work, it's easy to assume that every recurring task deserves a process around it. Sometimes it does. Sometimes “I need another screen” really can just be “I need another screen.”

Not Everything Needs to Be Automated

A few months ago, I was spending quite a bit of time building automations around hiring. Candidate details moving into the tracker. Reminders being triggered automatically. Information showing up where it needed to without someone having to copy and paste it every time.

I enjoyed building it. There was something satisfying about seeing a task that used to take ten minutes happen by itself. And, naturally, once one part was automated, I started looking at the next one.

At some point I asked myself:

“Why don't we automate the first message to candidates too?”

Technically, we could. Someone applies, the workflow picks it up, and a message goes out. Same response, same speed, no one has to remember to send it. It would have worked. But the more I thought about it, the less I liked the idea.

That first message is often someone's first interaction with WriftAI. Before they know our product, our team, or how we work, that's the first bit of the experience they have with us. Turning that into another automated step felt like we'd gained a little efficiency at the expense of something more important. So we left that part alone.

The chain stopped where it was needed, not where it could.

I'm still very much in favour of automating the boring parts. If a computer can move information between two places faster than I can, I'd rather let it do that. But not every part of a workflow is just a task waiting to be automated. Some parts are better when a person is still on the other side.

Some Meetings Have an Expiry Date

There was a period where Abdullah started most mornings with the same team. They'd sit down, talk through what was happening, what needed attention, and where someone might be stuck. It wasn't a particularly formal meeting. It was just a useful way of staying close to the work.

For a while, that was enough. Then the workload picked up. There was more happening across the team and more happening elsewhere too. The daily meeting started to become less about solving something together and more about making sure everyone had shown up to the same conversation.

So we gradually stopped doing it every day.

The team still gets together once a month, and Abdullah is still close to the work, but there's no longer a reason for everyone to spend time together every morning.

What I find interesting about this is that nothing was really wrong with the meeting. It had simply stopped being as useful as it once was.

I think that's an easy thing to miss as WriftAI grows. A way of working can be completely right for one stage and completely unnecessary at another. There's a tendency to treat anything we've introduced as something we now have to keep.

We're getting a little better at letting things expire. Not because they stopped working, but because they already did their job.

We Didn't Need a Rule for That

One of the first things I noticed when I started handling people-related decisions at WriftAI was how often I wanted to run things past Abdullah.

Someone had an issue; I'd bring it to him. Someone needed something, I'd ask what he thought. Even when I already had an idea of what I wanted to do, I'd usually check with him before doing it. It felt natural. People-related decisions can be sensitive, and having someone more experienced to talk things through with was useful.

Then I started noticing something.

I'd bring an issue to Abdullah and, instead of telling me what to do, he'd ask:

“What do you think we should do?”

At first, I still wanted his answer. I'd already been given the space to make the decision, but I was used to having the conversation before doing anything. Eventually I realised that was the point. He wasn't asking because he didn't know what to do. He was asking because he trusted me to decide.

So I started making the call myself.

Whether it was something around leave, a team concern, an internal request, or something that needed to be handled quickly, I didn't need Abdullah involved in every step. I'd keep him in the loop afterwards, but the decision was mine to make.

We could have formalised that. We could have written down which decisions needed approval, created thresholds for what I could handle myself, or added another process around people-related decisions. We didn't. Instead, the responsibility grew with the trust.

That's something I've come to appreciate pretty quickly at WriftAI. Not every increase in complexity needs another rule. Sometimes the better answer is simply to give someone more room to use their judgement. The process didn't scale. The trust did.

What Actually Needs to Scale

None of this is to say that we don't need processes, systems, or automation. We do. Some things at WriftAI have to get better as we grow. The product needs stronger foundations. The infrastructure needs to handle more. The way we work with customers' needs to stay reliable as more people depend on it.

What I've become more conscious of is that growth doesn't automatically make everything more important. A process that worked for five people might need changing when there are fifty. A manual task might eventually be worth automating. A conversation that once needed everyone in the room might not anymore. But the answer isn't always to add something simply because WriftAI has grown.

We're still figuring out that balance ourselves. I suspect we'll get some of it wrong, build things we later decide we don't need, and probably keep a few things around longer than we should. That's okay. For me, the important thing is that we're willing to question them.

If something helps us do better work, we'll keep it. If it doesn't, I don't think growth is a good enough reason to keep it around. If this sounds like the kind of place you'd enjoy working, you know where to find us.