AI can write the email. It can summarise the meeting, analyse information, generate campaign ideas, draft social content and turn one piece of material into ten different formats.
But there is one rather important thing it can't do for you
Decide whether the process was worth doing in the first place.
That distinction matters because businesses are understandably looking at AI as a way to become faster, leaner and more productive.
The danger is that we automate the wrong things.
A bad process completed in five minutes instead of five hours is still a bad process.
AI should start with the business problem
My interest in AI comes from marketing rather than technology.
I've spent more than two decades working across marketing, sales, business development and consultancy, helping organisations attract customers, develop new channels, launch services and turn ideas into something commercially useful. More recently, that has extended into practical experience with AI, automation and prompt engineering alongside digital marketing.
That background influences how I approach AI
I don't start with: What can AI do? I prefer: What are we trying to improve? Those are very different questions. The first encourages businesses to find uses for a technology. The second encourages them to find the right technology for a business problem. An AI prompt engineer still needs to understand the problem There is a lot of interest in the role of the AI prompt engineer, and understandably so. The quality of the instruction you give an AI system can dramatically affect the quality of what comes back. I've formally studied prompt engineering alongside AI, automation, content creation, SEO, analytics and digital marketing. But the more useful lesson has been that a good prompt isn't simply about finding clever words to type into a box.
You need context
What is the objective? Who is the audience? What information does the system need? What constraints should it work within? What does a good result actually look like? How will somebody check whether the answer is correct?
That sounds suspiciously like writing a good marketing brief.
And perhaps that's the point.
Prompt engineering is partly about communicating clearly enough for a machine to understand what you want. Good management has always required communicating clearly enough for people to understand what you want.
Neither works particularly well if you don't know what you want yourself.
Digital marketing doesn't need more content for the sake of it
This is particularly relevant to digital marketing in Kent and elsewhere because generative AI has dramatically reduced the effort required to create content.
A business can now produce blogs, emails, adverts, social posts, images and campaign ideas at a speed that would have been unimaginable a few years ago.
But customers haven't suddenly acquired more hours in their day to consume it all.
Producing more content isn't necessarily the opportunity.
Producing more relevant content, more efficiently might be.
My own marketing experience spans social media, email, Google Business, websites, partner channels, customer acquisition and B2B marketing. The lesson from working across those channels is that each needs a purpose.
AI doesn't change that.
If the audience is wrong, AI can reach the wrong audience faster
If the proposition is weak, AI can produce 50 versions of a weak proposition.
If the customer journey is confusing, an AI-generated advert can simply deliver more people into a confusing journey.
And if nobody knows what success looks like, AI can generate an impressive amount of activity without creating much value.
Automate the friction, not the thinking
There are plenty of marketing tasks where AI and automation can be genuinely useful.
Research, summarisation, first drafts, content repurposing, analysing information, identifying patterns, developing variations and handling repetitive administrative work can all become quicker.
That creates capacity.
What matters is what we do with it
If AI saves a marketing team five hours, the objective shouldn't automatically be to fill those five hours producing another 50 social posts.
Perhaps those five hours would be better spent talking to customers, analysing campaign performance, improving the website journey, developing partnerships or working out why people aren't converting.
The value of automation isn't the work it creates. It's the unnecessary work it removes.
Sometimes you should fix the process before you automate it
Before introducing AI into a workflow, I think it's worth looking at the workflow itself.
Take something as ordinary as a customer enquiry.
If an enquiry currently arrives in the wrong inbox, gets manually forwarded, requires somebody to find information held somewhere else, receives an inconsistent response and then isn't recorded properly, AI might help write the response.
But it hasn't solved the problem
The better questions are Why is the enquiry arriving there? Why does it need forwarding? Why isn't the information readily available? Why are responses inconsistent? Why isn't the outcome being captured? Fix those things first and then ask where AI can make the improved process faster. That's process improvement rather than technology adoption, and the distinction can save businesses considerable time and money.
AI needs human judgement
One of the things I've learned from working across start-ups, commercial organisations and public services is that technology changes much faster than the fundamentals of good business. Customers still need a reason to choose you. People still need clear objectives. Projects still need ownership.
Marketing still needs to produce an outcome
And somebody still needs to make a judgement about what matters. AI can contribute to that decision-making. It can interrogate information, challenge assumptions and surface possibilities that might otherwise be missed. But judgement still matters because an answer can be beautifully written, completely plausible, and wrong. Knowing when to question the output may ultimately prove more valuable than knowing how to generate it.
The opportunity for digital marketing in Kent For Kent businesses, particularly SMEs without enormous marketing departments, I think this is where AI becomes genuinely interesting. It can give smaller teams access to capabilities that previously required considerably more time or external resource. That can level parts of the playing field. But the competitive advantage won't last if everybody has access to broadly the same tools. The advantage moves elsewhere: understanding your customer better, asking better questions, developing a stronger proposition, making better decisions and using the technology intelligently.
That's why I see AI prompt engineering and digital marketing as complementary skills rather than a replacement for marketing experience
Knowing how to use the tool matters.
Knowing why you're using it matters more.
Don't automate something just because you can. Before using AI to automate or accelerate a marketing process, I'd ask five things
- What problem are we actually solving?
- Does this process need automating, or redesigning?
- What will AI improve: cost, speed, quality, insight or customer experience?
- Where does human judgement still need to sit?
- How will we know whether the change has produced a better business outcome?
