As a digital marketing director, I’ve learnt to treat generative AI tools the same way I treat any new channel: with curiosity, a test-and-learn mindset and a healthy dose of scepticism. Right now, three names come up in almost every client conversation – Claude, ChatGPT and Gemini. Each has genuine strengths; each has limitations and none of them should be mistaken for a replacement for experienced marketers.
ChatGPT remains the most versatile “all-rounder” in day-to-day agency life. It’s strong at quickly generating first drafts, pressure-testing messaging, producing campaign variations and helping teams think through structure – whether that’s a landing page, a comms plan, a content calendar or a pitch narrative. When you need a speedy starting point, or a sounding board that can iterate at pace, ChatGPT tends to be the tool people reach for first. The value isn’t that it produces perfect copy – it rarely does on the first pass – but that it accelerates the blank-page stage and improves throughout when time is tight.
Claude, in my experience, shines when the work demands clarity, tone control and careful handling of longer documents. It’s often particularly useful for summarising lengthy reports, turning meeting notes into readable outputs, or sense-checking whether a piece of writing actually flows for a human reader. When you’re working on thoughtful, brand-sensitive content – the kind that can’t sound generic or overly “marketing-y” – Claude can be a strong partner in shaping drafts into something more natural. It’s still a tool, not a voice of wisdom, but it can reduce the time spent wrestling with structure and repetition.
Gemini’s main advantage is context and integration. For teams operating heavily in Google Workspace, it can be useful for turning existing materials into usable marketing outputs – summarising documents, supporting presentation prep and helping translate messy internal inputs into something coherent. It also tends to be a helpful option when you want a tool that can work across multiple formats quickly, especially if you’re already living in Gmail, Docs and Sheets. For marketers, that matters because the job isn’t just “writing a post” – it’s pulling together insights, stakeholders and assets at speed.
So why does this need monitoring for client work? Because the value and risk profile changes quickly. These tools are improving month by month, new features appear constantly, and the ways people use them – and misunderstand them – evolve just as fast. Clients are increasingly asking: “Can you use AI to do this faster?” or “Why can’t we just generate everything ourselves?” If we’re not actively testing, we lose the ability to advise with credibility. And if we’re not setting guardrails, we expose clients to avoidable risks.
Those risks are real. AI tools can hallucinate facts, misinterpret context, and generate outputs that sound plausible but are wrong. They can also flatten brand voice into generic copy or unintentionally reproduce bias and stereotypes. From a compliance point of view, there are also questions around data handling, confidentiality and what should never be input into a model. In a communications environment where trust is hard won and easily lost, these are not theoretical concerns.
That’s why the right framing is “tool, not replacement”. The role of marketing experts doesn’t disappear – it becomes more important. Human judgement is how what to say, what not to say and what’s true is decided. Strategy is what turns outputs into outcomes. Creativity is what makes work distinctive. And relationship management – understanding stakeholders, reading a room, navigating sensitivities – is something no model can do.
Used well, Claude, ChatGPT and Gemini can save time, widen thinking and improve consistency. But the best results come when experienced marketers stay in the driving seat: setting the brief, reviewing the output, verifying claims and ensuring the final work is aligned to the brand, the audience and the moment. The winners won’t be the teams that “use AI instead of people”. They’ll be the teams that use AI to give people more space to do the high-value work only humans can do.