The Impact of AI on Agency Work: How We Use It

Let's start with the thing everyone worries about. AI hasn't replaced designers, developers, or strategists on our team, and we don't expect it to. What it has replaced is a lot of the tedious, repetitive work that used to sit between us and the actual creative and technical work we're here to do.

The useful mental model is that AI is a fast, tireless assistant, not an author. It drafts, summarises, and accelerates, but the judgment, taste, and final decisions stay with us. A tool that writes a first draft in seconds is only valuable if someone with real expertise shapes what comes next.

The hype tends to focus on flashy demos. The real value, for us, is in the quiet, practical things that save an hour here and an hour there, every single day. A few examples from our actual workflow:

Meeting notes and summaries. Client calls and internal meetings used to mean someone half-listening while scribbling notes. Now we let AI handle transcription and summaries, so everyone can be present in the conversation and walk away with a clean record of decisions and action items afterwards.

Overviews and research. When we're getting up to speed on a new client's industry or a technical topic, AI gives us a fast first overview to build on. It's a starting point for our own research, never the final word, but it shortens the path from "we know nothing about this" to "we're asking the right questions."

Drafting and summarising. First drafts of proposals, emails, and documentation come together faster when AI handles the blank-page problem. We use our Claude Pro subscription for a lot of this, drafting, restructuring messy notes, and turning long documents into clear summaries, then we edit heavily to make sure it sounds like us and says what we actually mean.

Automations. Small, repetitive tasks that used to eat time, formatting, reformatting, sorting information, repetitive copy variations, are increasingly handled by simple automations. Each one saves only minutes, but they add up across a week into real time we can spend on client work instead.

Just as important as where we use AI is where we deliberately don't. Creative direction, design taste, strategic decisions, and the final quality check stay firmly human. AI can suggest, but it doesn't understand a client's brand the way a person who's spent hours with them does, and it doesn't have the accountability that comes with putting your name on the work.

We're also careful with client confidentiality and accuracy. AI output gets checked, sensitive information is handled thoughtfully, and nothing goes to a client without a human having verified it. A tool that's right most of the time still needs someone responsible for catching the times it isn't.

The honest summary is this: AI hasn't changed what we do, it's changed how much of our time goes to the parts that matter. Less time on transcription, formatting, and first drafts means more time on design, strategy, and the relationships that actually define good agency work.

That's the shift we'd encourage other teams to think about too. The question isn't "how do we use AI to do more, faster?" It's "which parts of our work are busywork, and which are the craft?" Hand the first to AI, protect the second fiercely, and the tool earns its place.