AI POLICY
Be More, Human.
We use AI extensively. We have also spent a considerable amount of time testing approaches and working out where it helps and where it quietly damages the work. We tested the workflows, compared the output, and made our choices. What follows is not a position we are still forming. It is how we already work. We continue to evaluate developments in what is possible as they emerge, but there are certain principles we will not compromise on.
What we decided:
Judgment stays human
Strategy, positioning, architecture, and any decision a client is relying on us for is made by an experienced person who is accountable for it.
AI does the volume, not the deep thinking
Research, drafting, refactoring, and repetitive build work are where it earns its place. Those are real savings, and we pass them on.
Nothing ships unverified
Every AI-assisted output is reviewed by the specialist responsible for it. Facts are checked. Every line is read and tested.
Your information stays yours
We do not put proprietary material, client data, or unreleased product detail into platforms that would train on it.
We will tell you where it was used
Ask on any piece of work and you will get a straight answer.
Where we draw the line
We do not ship software, content or materials built by prompting alone. We do not use generated copy as final copy for positioning, technical content, or anything that carries your credibility. A model can describe what your product does. It cannot sit in the room where the hard decision was made, and your buyers can tell the difference.
This will change
The tools move quickly and so do we. What will not change is the part we are not willing to automate.