Ask us if we use AI and we’ll tell you the truth: yes, regularly. Ask us if AI picks the concept you’ll stake a rebrand on, and the answer is just as plain: no, never. Most agencies dodge the first question or overcorrect on the second. We’d rather answer both, and show exactly where the line sits in between.
The honest answer matters because the dishonest one has become the default pitch. Some shops market themselves as “AI-free,” which is either untrue or a competitive disadvantage they’re volunteering for. Clients already see the tools their own staff use every day; a vendor who claims total abstinence has either not noticed or isn’t being straight. Others let a generator make calls a human should be making and call it efficiency. Neither position survives a direct question about process. Ours does, because we drew the line somewhere specific — not somewhere convenient.
What AI Actually Does in Our Process
AI earns its place in the parts of the work that are mechanical, repetitive, or exploratory — the parts where speed doesn’t cost us anything we’d otherwise be selling.
- Competitive and category scans. A fast first read on how twenty competitors in a category position themselves, before a strategist spends real hours confirming or correcting it.
- Early moodboard and typographic exploration. A wide first pass of directions, so a designer starts from forty options instead of four.
- Production variants. Resizing, reformatting, and localizing an already-approved system across channels and markets — pure execution, once the creative decision is made.
- First-pass internal drafts. Early versions of internal documentation or brief language that a strategist rewrites before anything reaches a client.
Run that research step without a person in the loop and here’s what comes back: twenty competitor positioning statements that all say some version of “innovative, trusted, mission-driven.” Useful raw material. A real problem if anyone mistakes it for analysis instead of a starting point. The strategist’s job is exactly the part the tool skipped — deciding what any of it actually means.
Each of those four uses speeds up a person’s work. None of them replaces the judgment call sitting at the center of it.
Where the Market Agrees With Us
We’re not alone in drawing this particular line, and recent data backs up why. In a May 2026 study of 1,415 marketing leaders across seven countries, conducted with The Harris Poll, Canva found that 97% of marketers already use AI in their everyday creative work. The same study asked those leaders what AI still can’t replicate. The top answers: empathy and emotional intelligence, the human imperfection that sparks originality, and brand intuition and creative judgment — each cited by roughly four in ten respondents.
That’s not agencies protecting their own billable hours. That’s people running AI-saturated marketing organizations, naming the exact things we refuse to automate. Consumers are drawing a similar line from the other side of the same study: 87% say the best advertising still requires a human touch, and 78% said they’d rather see an ad made by a person, even if AI could technically make a better one.
The Decisions We Never Hand Off
Four calls stay with a person, every time, regardless of deadline or budget pressure:
- The positioning itself. Where a brand sits relative to competitors is a bet on the business, not a creative output. A model has no stake in whether that bet is right.
- Which concept wins. We generate wide, but a person narrows. That decision weighs context — competitive pressure, sales feedback, internal politics — none of which makes it into a prompt.
- The rationale we defend in the room. When a client pushes back, someone has to explain why a direction is right and mean it. A generator isn’t in that meeting and has nothing on the line if it’s wrong.
- Anything touching legal or cultural risk. Trademark exposure, cultural connotations in a new market, accessibility failures — we’ve covered what AI alone misses here before, and the short version hasn’t changed: judgment with something at stake beats a model with nothing to lose.
A positioning decision that’s wrong doesn’t surface in a design review. It surfaces eighteen months later, in a sales team that can’t explain why deals keep slipping to a less interesting competitor. That’s the real cost of outsourcing judgment: it’s invisible right up until it isn’t. Nobody traces that back to a prompt. Everybody should.
The Line Moves. Our Standard Doesn’t.
The tools will keep getting better at the mechanical half of this job, and when they do, we’ll use more of them there. What won’t change is the half we don’t delegate: the call on who you are, which direction is actually right, and who’s willing to defend that answer when it’s challenged. A tool with no stake in the outcome can’t be fired for being wrong, can’t be held to the brief, and won’t be in the room when a board asks why a decision got made. We will. That’s not a stance against AI. It’s a stance for having someone accountable in the room — which, lately, feels like the rarer commitment.
If you want to know exactly where a partner draws that line before you sign anything, that’s a fair question for a first conversation — not something you should find out after launch. Get in touch.
