Ask a digital agency whether they use AI and you'll usually get a nervous pause, followed by either a proud yes or a defensive no. Both miss the point. AI is already in almost everything you read, click and buy online, so the honest question was never whether anyone uses it. It's how.

That's the whole game now: AI assisted vs AI generated. One uses the technology to make skilled people faster. The other hands the work to a machine and hopes nobody notices. We sit firmly in the first camp, and we think you should expect the same from anyone you trust with your brand, your product or your customers.

To be honest, the line between the two is easy to say and harder to hold. So here's exactly where we draw it, why it matters, and what "AI assisted, not AI produced" means for the work we do for you.

‍

The question isn't whether you use AI. It's how

Here's the reality most agencies won't say out loud. AI is already baked into the web. According to Ahrefs, 74.2% of newly published web pages now contain some AI-generated content, and only a tiny fraction of those are pure machine output. The rest are a blend of human and AI work.

It holds true at the top of the quality scale too. The same research found that 86.5% of top-ranking pages contain some AI-generated content, with effectively no link between AI use and any ranking penalty.

Google has been clear about why. It judges content on whether it's helpful, accurate and original, not on whether a person or a tool produced the first draft. There's no penalty for using AI. There's a very real penalty for publishing thin, generic work, however it was made.

This matters because the risk has quietly flipped. A few years ago the worry was using AI at all. Today the bigger risk is using it badly: shipping fast, generic work at scale and watching it drag down the quality, trust and search visibility you've worked hard to build. The tool was never the problem. Reaching for it instead of thinking is.

So "do you use AI?" is the wrong question to ask your agency. The one that actually tells you something is "how do you use it, and who's accountable for what you ship?"

‍

What "AI assisted, not AI produced" actually means

The short version: AI assisted means a person owns the work and uses AI to get there faster. AI produced means the machine writes it and someone pastes it in. We do the first. We never do the second.

In practice, AI helps us with the heavy lifting. Research, first drafts, repetitive code, asset variations, the early scaffolding of an idea. It gets us to a strong starting point in a fraction of the time it used to take.

Then the real work begins. Our designers, developers and strategists shape it, challenge it, fact-check it and make the calls that need taste and context. The thinking is ours. The judgement is ours. The final word is always a human one.

Picture how a single page or feature actually moves through our studio. AI might pull the research and rough out a first structure in an afternoon. From there a writer rebuilds it around a real point of view, an engineer pressure-tests anything technical, and an editor cuts the filler the model loves to add. By the time you see it, very little of that first machine draft survives, and what's left has earned its place.

That's the rule we hold on every project: we never just cut and paste. If something leaves our studio with your name on it, a person has read every line, stands behind it, and can tell you exactly why it's there.

‍

Why AI produced work falls over

Raw AI output has a habit of sounding fine and saying nothing. By design, these tools predict the most likely next word based on everything already out there, so left alone they produce a polished average of the internet. Confident, competent, and completely forgettable.

The people doing this for a living know it. Across the industry, the overwhelming majority of teams edit or heavily rework AI drafts before anything reaches a customer, and very few rate the quality of AI-only work as good enough to stand on its own. Nobody serious is shipping the raw stuff on purpose.

There's a quieter cost too. When everyone leans on the same handful of tools with the same prompts, the work starts to converge. Pages blur into one another, brands lose the edges that made them recognisable, and the voice you spent years building ends up sounding like everybody else's. Sounding like the average of your competitors was never the goal.

Your customers feel it even when they can't name it. As these tools have spread, trust in online content has been sliding, and a wall of obviously machine-made copy is a fast way to feel less human to the people you're trying to win over. That kind of trust is hard to win back once a brand starts to feel automated.

Then there are two deeper problems that no amount of clever prompting fixes. The first is experience. AI can describe a thing it has never done, but it can't bring the scar tissue of having actually built, launched or rescued something. That lived knowledge is exactly what makes work worth reading and worth trusting, and it's the heart of keeping a human in the loop.

The second is accountability. When a machine invents a statistic or a source, and it does, someone still has to answer for it. A person can own a mistake and fix it. Software can't. Put your name to AI produced work and you're vouching for something nobody actually checked.

‍

Where AI earns its place in our workflow

None of this means we're precious about the tools. Used well, AI is one of the most useful things to happen to how we work in years. The trick is pointing it at the right jobs and keeping people on the ones that count.

‍

The jobs we happily hand to AI

Some work is repetitive, time-consuming and adds little when a person does it by hand. That's where AI pulls real weight for us:

Take a website audit. AI can scan a site and surface dozens of performance, accessibility and SEO issues in minutes. What it can't do is tell you which three of those will move the needle for your business, or how to fix them without breaking something else. That judgement is the part you'd never want a machine making on its own.

The same goes for code and design. AI will happily generate a component or a layout, but it doesn't know your customers, your edge cases or the corner where a clever shortcut becomes next year's headache. We use it to clear the obvious so our people can spend their time on the decisions that actually carry risk.

‍

What real AI workflow efficiency looks like

This is what we mean by efficiency meets experience. AI gets us to a strong starting point faster. Our people take it the rest of the way. You get work that's quicker to produce without losing the things that make it good.

Done right, the time we save on the groundwork goes straight back into the work that needs a human. More thinking, more craft, more rounds of getting it right. That's the only kind of AI workflow efficiency worth having, and it's the opposite of using the tools to quietly do less.

‍

Building AI into client products, carefully

The same thinking shapes how we build AI into the products we make for you. Where it solves a real problem, we'll integrate it properly. Where it's there to tick a trend box, we'll tell you to save your money.

‍

Start with a real problem, not a trend

Good AI features start with a genuine need. A faster way for your customers to find what they want, a smarter search that understands what people mean rather than just what they type, a support tool that actually helps rather than frustrates. We design these for accuracy, safety and the people using them, not for a line in a press release.

It's just as much our job to say no. If a feature won't earn its keep, or it adds risk your customers will end up paying for, we'd rather tell you straight than ship something shiny that quietly works against you. The best AI decision we make on a project is sometimes leaving it out altogether.

‍

Responsible AI means owning the outcome

Responsible AI in any product also means being careful with data and honest about limits. We think hard about what information goes into these systems, we keep your customers' trust front of mind, and we make sure a human stays accountable for what the product does once it's live.

No AI feature ships from us as a black box nobody understands. We build in the checks, we test how it behaves when things go wrong, and we make sure there's always a person responsible for the result. Speed is good. Speed at the cost of trust never is.

‍

The bottom line

If you take one thing from this, make it the question. Not whether your agency uses AI, but how, and whether a person is accountable for the result. That single distinction, AI assisted versus AI produced, is what separates work you can rely on from work that just fills space.

Our position is simple. AI does the lifting. Our people own the outcome. That's how you move fast without giving up the craft, the experience and the accountability that make work worth paying for.

If you'd like a partner who uses AI to make good work faster rather than to cut corners, we should talk. You can see how we work on our services page, get to know us on our about page, or read our full AI policy to see exactly where we stand.

‍

What's the difference between AI assisted and AI generated content?

AI assisted content is created by a person who uses AI to work faster, then shapes, checks and signs off the result. AI generated content is produced by the tool and published with little or no human input. The first keeps human judgement and accountability. The second hands both to a machine.

‍

Does Google penalise AI generated content?

No, not for being AI generated. Google judges content on whether it's helpful, accurate and original, not on how it was made. What it does filter out is thin, generic, low-value work, which happens to be exactly what unedited AI tends to produce.

‍

Is AI assisted content better than AI generated content?

For anything that carries your name, yes. AI assisted work brings real experience, original insight and a person who's accountable for it. It's also why almost nobody publishes raw AI: work shaped and checked by skilled people simply lands better with readers and search engines alike.

‍

Will AI replace agencies and creative teams?

We don't think so, and the work tells us why. AI is brilliant at speed and scale, but it has no taste, no lived experience and no accountability. The agencies that thrive will be the ones using AI to make their people better, not the ones trying to replace them.

‍

Should businesses disclose when they use AI?

Being open about how you work is rarely a bad thing, and we're happy to tell clients where AI fits into their project. What matters most to your customers, though, is that the end result is accurate, useful and genuinely yours. Honesty about the process and quality in the output go hand in hand.