We use AI.
That sentence should probably not be controversial anymore.
The more useful question is how.
At intrstd, we do not see AI as a separate service or a magic layer that gets added to a project because it sounds contemporary. We treat it like any other tool in the studio: useful when it improves the work, unnecessary when it does not.
Our position is simple: automate effort where it helps. Do not outsource judgment.
We use AI to understand faster
Projects often begin with a lot of material and very little clarity.
Old websites. Presentations. Meeting notes. Customer feedback. Competitor pages. Product sheets. Emails. Half-finished briefs.
AI can help organise and summarise this material so we can get to the useful questions faster.
It can show repeated themes. Surface inconsistencies. Turn transcripts into structured notes. Compare what the sales deck says with what the website says.
What it cannot know is which contradiction matters politically, strategically or commercially. That remains a conversation.
We use AI to explore
Early creative work benefits from breadth.
We may use AI to test visual territories, image directions, content structures, headline alternatives, naming families or ways of expressing an idea.
This is exploration, not approval.
The fact that something can be generated does not mean it becomes client work. Most of it will not.
We use AI for roughness
Oddly, one of the best uses of AI is making something concrete enough to discuss before spending too much time finishing it.
A rough packaging world. A possible hero image. A quick prototype. A placeholder illustration. A version of the copy that helps everyone understand the intended hierarchy.
These are thinking aids.
They let us ask “Is this direction right?” before asking “Is this execution perfect?”
We use AI for production
There are jobs where doing them manually does not make the outcome more thoughtful.
Background removal. Image extension. Resizing. Drafting alt text. Organising files. Transcribing meetings. Creating first-pass summaries. Checking repetitive content. Generating test data. Exploring responsive variations.
When the tool can do these reliably, we use it and review the result.
That last part does not disappear.
Where we deliberately keep humans in charge
Positioning. What should this business stand for? What problem are we actually solving? What should change in the mind of the customer?
Brand direction. We may use AI during exploration, but the core identity is selected because we believe it is the right response to the brief.
Client conversations. We want to hear what people mean, not just what was written in the brief. Hesitation, disagreement, internal politics and enthusiasm are information too.
Final copy. AI can help draft and edit. A human has to decide whether the claim is true, whether the tone is right and whether the words actually sound like the business.
Final approval. A human takes responsibility for what leaves the studio.
We do not sell AI as a shortcut around understanding
There is a temptation to imagine that better prompting can replace discovery.
It cannot.
If the business does not know who it is for, what it does differently or what customers value, asking a model to create the brand strategy simply produces a confident interpretation of incomplete information.
Sometimes that interpretation is useful as a provocation.
It is not the same thing as knowing.
We also do not hide it
If AI materially contributes to a piece of work, we are comfortable saying so.
There is no prize for manually doing a task that a tool can perform better or faster.
The standard is not whether a machine was involved.
The standard is whether the final work is original enough, accurate enough, appropriate enough and well made enough to carry our name.
The point is better work
AI will continue changing. Some tools we use now will disappear. Other capabilities will become ordinary parts of software we already use.
The studio principle does not need to change with every release.
Think first. Use the right tool. Check the result. Keep the human responsible.
AI is part of the process. It is not the point of the process.