Why we refocused

For a while, Gruow did a bit of everything: websites, brand identities, product design, development, training, audits. If someone needed it and we could do it, we said yes.

That made sense at the start. Saying yes is how a small studio finds work and finds out what it’s good at. Over time, it cost more than it gave back.

What doing everything cost us

It was hard to explain. When someone asked what Gruow does, the honest answer was a list. A list is easy to agree with and hard to remember. Someone whose advice we trust put it plainly: there was no clear reason to pick us over anyone else.

It was hard to get good. Depth comes from doing the same kind of work again and again, and seeing what holds up. Every new kind of job meant starting close to the beginning.

It hid the work we do best. The work we kept coming back to was building things that make work run: tools, plugins, and small pieces of software that get two systems talking. That work was buried somewhere in the list.

What changed

AI changed what software can do for a business, and how quickly it can be built. The models are now good enough for most of what businesses would ask of them. The open question is where they fit in a company’s work, and how to build them in so the work actually gets easier.

That question doesn’t belong to strategy, design or code alone. It needs all three, on the same piece of work, by the same people. That’s what we want to be good at.

What we do now

Strategy, design and code, applied to three kinds of work:

• Products: new software, built from the idea through to the first working version.
• Internal tools: software a business runs on, built around how its people actually work.
• Integrations: connections between the systems a business already has, including MCP servers that let AI tools work with its own data.

Strategy is always tied to something we build. We don’t make mobile apps or no-code workflows, and we don’t maintain systems we didn’t build.

Learning as we go

We don’t have this figured out, and nobody does yet. The tools change month to month, and so does what works.

So we research as we work: new models, new tools, and how teams actually take them up. We publish what we find here. Some of it shapes what we build. Some of it will turn out wrong, and we’ll say so when it does. Our own products, E10s and Hallful, are where we try ideas on ourselves first.

Focus doesn’t mean doing less. It means each piece of work teaches us something we use on the next.