There's a particular kind of agency pitch doing the rounds at the moment. It goes roughly like this: "We use cutting-edge AI to supercharge your marketing results." What follows is usually a vague description of using ChatGPT to write content, some stock imagery that isn't quite right, and a pricing model that somehow costs more than the same work done by a competent human.
The honest picture of AI in marketing is more nuanced and more interesting than either the breathless hype or the reflexive scepticism. There are places where AI genuinely changes what's possible for marketing teams — and places where the hype significantly outstrips reality. Here's an attempt at an honest map of both.
Where AI Is Genuinely Useful
Research and synthesis at speed
One of the most genuinely transformative uses of AI in a marketing context is research. Summarising competitor positioning from dozens of sources, synthesising customer review data across hundreds of reviews, identifying themes in large bodies of qualitative feedback — tasks that would take a human analyst days can be completed in minutes with the right AI tooling.
This isn't about replacing insight; it's about compressing the time it takes to gather the raw material for insight. A strategist who previously spent two days reading competitor websites can now spend two hours doing the same thing and use the remaining time thinking rather than reading.
First drafts and copy variants
AI is a useful first-draft engine. Not a final-draft engine — but a first-draft engine. For tasks like writing ten subject line variants for an email, generating initial copy for a Google Ads campaign, or producing a rough first cut of a product description, AI is genuinely faster than a human copywriter starting from a blank page.
The catch: AI-generated copy is typically average. It regresses to the mean of everything it's been trained on. For commodity copy tasks where "adequate" is sufficient, that's fine. For copy that needs to carry a distinctive brand voice, make a nuanced argument, or connect emotionally with a specific audience, AI output typically requires significant human editing to be usable — and in many cases, starting from scratch is faster.
The right mental model: AI as a junior copywriter who can produce quantity quickly but requires careful oversight and editing.
Personalisation at scale
Dynamic personalisation — different content for different audience segments, triggered by behaviour or data signals — has been theoretically possible for years but practically difficult because it requires enormous amounts of content production. AI is beginning to change that equation by making it feasible to produce personalised variants of email sequences, landing pages, and ad creative at a scale that human teams couldn't manage.
This is one of the more commercially significant AI applications in marketing, and it's still relatively early. Businesses that get this right over the next two to three years will have a meaningful performance advantage.
Data analysis and pattern recognition
Marketing generates data. Large amounts of it. AI tools are increasingly useful for identifying patterns in performance data that human analysts might miss — which audience segments respond best to which creative approaches, where in the customer journey drop-off is occurring, which channels are generating leads that convert at higher rates downstream.
This isn't a replacement for human judgement about what to do with the patterns — but it does surface them faster and more reliably.
Image and visual generation (in specific contexts)
AI image generation has reached a quality threshold where it's genuinely useful for specific tasks: mood board creation, rapid concept visualisation, social media background images, and illustration-style content. For brand photography or product photography, AI is still not a substitute for a proper shoot. But for low-stakes visual content needs, it's a legitimate time-saver.
Where AI Is Mostly Hype
Replacing strategic thinking
Strategy requires context that AI doesn't have: deep knowledge of your specific business, your competitive dynamics, your customer relationships, your resource constraints, and your risk tolerance. AI can simulate strategy at a surface level, but the output is typically generic — a list of things that are broadly true of every business rather than specifically true of yours.
The businesses getting genuine value from AI in marketing are using it to accelerate execution, not to replace the thinking that precedes it.
"AI-powered" content at volume
The idea that AI can produce blog posts, articles, and website content at volume, cheaply, at a quality that will perform in search or build genuine audience trust — this is the biggest piece of marketing AI hype to be sceptical of.
Google's helpful content system is explicitly designed to surface content that demonstrates real expertise, experience, and genuine value to readers. The kind of content that performs consistently well in search is the kind of content that contains original insight, specific evidence, and a clear point of view — things that are difficult to produce at volume by AI and easy to produce by people who actually know what they're talking about.
Agencies promising AI-powered content volume as a path to SEO growth are, in most cases, creating a short-term numbers game that will cost their clients in quality signals and trust over time.
Customer service replacement
AI chatbots for customer service have improved significantly, but the bar of customer expectations has risen with them. A chatbot that fails to understand a nuanced question — or worse, confidently gives wrong information — damages brand trust in a way that's difficult to quantify but real. In most B2B and high-consideration B2C contexts, AI should augment rather than replace human customer service.
A Framework for Evaluating AI Marketing Tools
When you encounter a new AI marketing tool or an agency pitch featuring AI capabilities, ask three questions:
- Is this automating a task that was previously limited by volume or speed? If yes, it's probably genuinely useful.
- Is this producing output that requires human judgement to be valuable? If yes, factor in the cost of that human time.
- Is this a shortcut around something that matters — brand quality, strategic rigour, genuine customer insight? If yes, be sceptical.
AI is most valuable when it compresses time on tasks that are inherently high-volume, low-judgement, and well-defined. It's least valuable when it's positioned as a replacement for the things that actually differentiate excellent marketing from average marketing: insight, creativity, strategy, and genuine human expertise.
At Brainwave Designs, we use AI tools where they genuinely make our work better or faster — and we're transparent about where. If you'd like to talk about how to build an AI-augmented marketing function that delivers real results, get in touch.
"AI in marketing is a multiplier. It multiplies what you put in. If you put in mediocre strategy and generic briefs, you get mediocre results faster. If you put in genuine insight and clear direction, you get excellent results more efficiently."
Found this useful? Share it with someone who might benefit.
Back to Insights
