Why the best work still starts and ends with people
Let’s start with the concern: if a machine can write the copy and analyse the data, what exactly are clients paying an agency for? The honest answer is that AI does not replace the thinking that makes good work good. It reduces the friction between the thinking and the output. Our value was never in the typing itself. It has always been in the judgement, knowing what to say, to whom, and why. AI can help with execution, but it cannot provide that judgement.
Speed is the headline. Quality is the story
The obvious benefit is speed. Work that might previously have taken days or weeks can now be delivered in hours. But speed on its own is not the point, and faster bad work is still bad work. The real value lies in what we do with the time we get back: more refinement, more attention on strategy, and more opportunity to challenge whether the brief is right in the first place. The agencies that use AI well are not lowering the bar; they are raising it.
Human in, human out
A simple principle keeps this grounded: a person at the start, and a person at the end. AI is a highly effective processor of human input, but a vague prompt will still produce a vague result. The quality of the output depends heavily on the quality of the brief. Equally, nothing should reach a client without a person checking it for quality, relevance and accuracy. AI is often very confident, including when it is wrong. Spotting that remains part of the job. People set the direction, people review the outcome, and the technology does the heavy lifting in between.
Why transparency isn’t optional
Too many agencies still seem reluctant to admit they use AI at all, for fear that it somehow cheapens the work. We think that is the wrong approach. Trust, once undermined, is difficult and expensive to rebuild. Being open about how AI fits into the process signals confidence and control, and it shifts the conversation away from whether corners are being cut towards how rigorously and responsibly the work is being done.
There is, however, a sting in the tail. You can deliver something genuinely ambitious, explain that AI played a part, and still receive the rather flat response: “but it was done with AI.” What that overlooks is the work behind the work – the ideation, the segmentation, the integration, judgement, the iteration, and the trial and error required to get to a strong result. AI did not hand us the answer. It was simply one of the tools used skilfully to get there. So saying “we used AI” should never be taken to mean “we pressed a button.” If anything, the work has not become easier; it has become more ambitious.
AI is software, not magic
The clearest way to deal with that point is to reframe what AI actually is: software. On its own, it produces nothing of value. The outcome depends entirely on the skill and judgement of the person using it. We have seen this before. An accountant who builds a strong forecasting model in a spreadsheet is not dismissed as someone who merely let Excel do the work, and today that argument would sound faintly absurd. AI is likely to follow the same path. Before long, saying “you used AI” will be no more remarkable than saying you used a laptop. What will continue to differentiate agencies is exactly what always has: how well they use the tools available to them.
The things we couldn’t do before
It would be a mistake to position AI purely as a faster way to do existing work. More interestingly, it makes genuinely new things possible: testing twenty creative routes instead of two, personalising content at a scale that would previously have been unrealistic, identifying patterns across large data sets in minutes, or refining ideas live in a client conversation. Even something as routine as meeting notes can now be captured in far greater detail, with actions, nuances and offhand remarks transcribed and summarised so that important points are not lost. These are not shortcuts. They are capabilities that simply did not exist at this cost a few years ago, and they allow us to offer clients more ambition rather than less effort.
The bottom line
AI has not made expertise less valuable. If anything, it has made it more visible by removing some of the busywork that used to obscure it. The agencies that thrive will be honest about how they work, keep people firmly at both ends of the process, and use the time they save to improve the quality of the work rather than simply increase the speed of delivery. Put simply: quality in, quality out, and nothing to hide.
