AI in Agency Work
5 min read

How We Use AI in Our Agency (And Where We Don't)

Transparent breakdown of where AI genuinely helps in agency work and where it falls flat. Builds trust through specificity and honesty about the limits of AI.

I'll tell you exactly where we use AI. And where we refuse to.

This isn't a confession or a defence — it's just transparency. Because the question of AI in agency work is one every client is quietly asking, and deserves a direct answer.

Where AI Genuinely Helps Us

Research and synthesis. Briefing on a new sector, summarising competitor positioning, getting up to speed on a technical topic quickly — AI has meaningfully compressed the time this takes. What used to require several hours of reading can now yield a solid working brief in 30–45 minutes. The result still gets verified and supplemented with primary research, but the starting point is better.

Ideation. Generating article ideas, headline options, CTA variations, content structures. We don't use AI output directly, but we use it to generate volume quickly — 20 headline options from which we select and refine 2 or 3. It's a creative sparring partner that never runs out of suggestions and doesn't have an ego attached to its ideas.

First drafts for templated content. Meta descriptions, social post structures, email subject lines, FAQ answers — structured, repeatable content with clear constraints. AI handles the scaffolding. Humans refine the voice.

Code snippets and debugging. Specific, bounded development tasks: writing a custom JavaScript function, debugging a CSS issue, generating a regular expression. Significant time savings on tasks with objectively correct answers.

Documentation. Internal process documentation, client onboarding materials, project briefs. Structure-heavy, information-dense content where consistency matters more than voice.

Where AI Falls Flat

Strategic thinking. AI doesn't understand a client's business the way a strategist who's spent hours in their world does. It can describe positioning frameworks. It can't feel where a brand is stuck, identify what a leadership team is really afraid of, or sense where the market is moving before the data confirms it.

Strategy requires judgment built from experience and context. AI can inform it. It can't replace it.

Brand voice and tone. Generic AI content has a recognisable texture: confident, balanced, vaguely helpful, slightly corporate. It's not wrong, but it's not distinctively you. Every brand we work with has a voice — a way of talking to their audience that reflects their personality, history, and genuine point of view. That takes work to define and maintain. AI copies patterns. It doesn't create voice.

Emotional resonance. The writing that converts — the headline that stops someone scrolling, the case study that makes a prospect feel seen — comes from a specific kind of empathy. Understanding not just what the audience thinks but what they feel. What they're afraid of. What they want to be. AI processes language. It doesn't feel the gap between where someone is and where they want to be.

Relationship work. Client relationships, partnerships, complex conversations. AI can draft a message. It can't read a room, manage anxiety, or build the kind of trust that comes from showing up consistently over time.

The Honest Picture

AI has made our team faster at a specific set of tasks. It hasn't made us better at the things that actually differentiate good agency work from average agency work.

The agencies that will thrive in an AI-enabled world are the ones that use it to move faster on the commodity tasks — and reinvest that time into the irreplaceable ones. Better strategy. Deeper client understanding. More distinctive creative work.

The agencies that will struggle are those who don't use AI at all (falling behind on efficiency) or those who use it for everything (producing undifferentiated output at speed).

The edge is in the judgment about where AI adds value and where it doesn't. That judgment is itself something AI can't help you with.

What This Means for Our Clients

When we use AI in a project, we're using it on tasks where it genuinely accelerates quality work — not as a replacement for the thinking that makes the work worth having.

Your strategy is still developed by strategists. Your brand voice is still defined by writers who understand your business. Your website is still designed by people with aesthetic judgment and commercial instinct.

AI helps us work faster. You benefit from the time savings. You don't get lower-quality output — you get the same quality, delivered more efficiently.

That's the honest version. We don't think it needs apologising for.