How we work
Marketing is being remade right now, mostly by AI. Most teams put AI on top of a way of working that was shaped long before it existed. We build the way of working around the technology from the start, and this is what we've figured out so far.
AI first, not AI only
Much of marketing is production. Research, first drafts, setup, reporting. AI handles that work today, and it's exactly the work that used to fill a whole team's days.
What's left is judgment and craft. One experienced person making those calls goes much further than a few years ago, without the quality slipping.
What AI doesn't do
AI can summarize what a client said, but not understand what they meant. What happens between the lines of a conversation is still human.
Strategic decisions can't be delegated either. AI lays out the options, someone has to choose. And the client relationship is not a task you can automate away, it is the job.
Knowing what takes what
AI does one part of the work, a human another, and a specialist a third. Telling which is which takes experience, and the wrong call gets expensive.
That's why a senior person always owns that decision. A junior with AI produces ten times more, but ten times more of almost good still isn't good.
Context is the bottleneck
The model is rarely the problem. What decides the quality is what the AI has to work from: customer data, past work, tone of voice, what's already been tried. Garbage in, garbage out applies in full.
That's why we put as much time into building the context as into writing the prompts.
The bar has moved
What used to be good enough now drowns in everyone else's generic content. When production becomes free, the value is in having something to say.