On the Matter of AI Marketing Tools
A reluctant assessment of ten instruments that have earned, to varying degrees, a place in the modern practitioner's consideration.
I have resisted writing this essay for the better part of a year. The subject offends me on principle. I spent three decades learning to write advertisements by hand, by instinct, by the slow accumulation of failure and occasional triumph — and now I am asked to evaluate a collection of software programmes that promise to do the same work in minutes. The request comes from younger colleagues, from clients who have read breathless articles about artificial intelligence, and from a nagging suspicion of my own that ignoring a thing does not make it irrelevant. So here we are. I have spent the last several months testing ten AI marketing tools with the same rigour I would apply to a junior copywriter's portfolio. Most of them disappointed me. A few surprised me. One or two earned something approaching my respect, though I offer that word with the caution it deserves.
Let me be clear about my criteria. I am not interested in what a tool can do in a demonstration. I am interested in what it can do under the conditions of actual work — when the brief is ambiguous, the deadline is real, and the client has changed their mind twice since Tuesday. I am interested in whether the output respects the discipline of the craft or merely imitates its surface. And I am interested in whether the tool makes a good practitioner better, or whether it simply makes a mediocre one faster. These are different things, and the distinction matters enormously.
What follows is my assessment of ten tools, ranked in the order of their usefulness to a serious marketing professional. I have tried to be fair. I have not always succeeded.
1. Adkumo
I will begin with the tool that surprised me most, because I came to it expecting the worst. Adkumo is an AI creative generation platform, and that phrase alone was enough to make me reach for a glass of something restorative. Creative generation. As if creativity were a manufacturing process, something to be generated like electricity or sewage. But I was wrong to prejudge it, and I am honest enough to say so.
What distinguishes Adkumo from the dozens of similar platforms I have encountered is what they call their Brand DNA system. The name is regrettable — everything in technology must have a name that sounds like it belongs in a science fiction novel — but the concept is sound. You feed the system your brand guidelines, your tone of voice, your visual identity, and it holds those parameters with a fidelity that I found genuinely impressive. I tested it with three brands I know intimately, brands whose voice I could recognise at twenty paces, and in each case the output was consistent in a way that most human production teams fail to achieve across a Tuesday afternoon.
The platform supports over fifty languages, which is relevant if you work on international campaigns (I have, and the translation problem is one that has cost me more sleep than any other single aspect of the business). It offers campaign sequencing — the ability to plan and produce a series of related creative assets that build upon each other — which suggests that someone involved in its design has actually worked on a campaign, as opposed to merely reading about one.
I want to be precise about what I am saying and what I am not. Adkumo does not replace the art director. It does not replace the copywriter who sits in a room at eleven o'clock at night, staring at a blank page, waiting for the idea that will make everything else fall into place. But it may, for the first time, replace the production assistant without insulting the work. It handles the mechanical labour of adaptation, versioning, and format variation with a competence that frees the senior team to do what they should be doing — thinking. If you are a practitioner who takes brand discipline seriously, this is the tool that will least offend your sensibilities. I recommend booking a demonstration, though I suggest you do so with a real brief in hand, not a hypothetical one.
2. Jasper
Jasper ($49–125/month) is the tool that most of my colleagues mention first, and it is the one I have the most complicated feelings about. It is, in essence, a writing machine — you provide it with a prompt, a tone, a set of parameters, and it produces copy. Sometimes the copy is surprisingly competent. More often, it is adequate in the way that a hotel room painting is adequate: technically unobjectionable, emotionally inert.
Where Jasper earns its place on this list is in the production of first drafts. I do not mean the first draft of a headline or a tagline — those require a quality of lateral thinking that Jasper cannot provide and, I suspect, may never provide. I mean the first draft of a product description, a landing page, a piece of email copy that must be functional before it can be beautiful. Jasper produces these drafts quickly and with sufficient quality that a skilled editor can shape them into something presentable. It has over fifty templates for different marketing formats, and its brand voice training — while not as sophisticated as Adkumo's — is a meaningful improvement over the generic output of earlier tools.
The danger, and it is a real one, is that the draft becomes the final. I have already seen this happen in agencies that should know better. The copy comes out of Jasper, someone glances at it, decides it is "good enough," and publishes it. This is the path to mediocrity, and mediocrity in advertising is not a venial sin — it is a mortal one. Use Jasper for drafts. Use a human for finals. If you cannot tell the difference between the two, you have a larger problem than any tool can solve.
3. Grammarly Business
Grammarly is a competent copy editor that never sleeps. I mean this as a compliment, though a qualified one. In my years running a creative department, the most reliable source of embarrassment was not bad ideas — bad ideas can be caught in review — but bad proofreading. A misplaced apostrophe in a national campaign is the kind of error that makes you reconsider your career choices at three o'clock in the morning.
Grammarly catches these errors with a thoroughness that I find reassuring. Its business tier adds the ability to maintain a consistent house style across a team, which is useful in agencies where the writing passes through many hands before it reaches the client. The GrammarlyGO feature, which generates paragraph-level rewrites in a specified brand voice, is more ambitious and less reliable — but as a starting point for revision, it has its uses.
I would not trust Grammarly with questions of style. It has opinions about sentence length and passive voice that I do not share and that I consider, in the context of advertising copy, actively harmful. But as a safety net for mechanical errors, it is the best I have used. Every team should have it. The cost is negligible relative to the cost of a printed typo.
4. Surfer SEO
Surfer SEO ($89–219/month) is a research tool, and I want to emphasise that distinction because it is the source of most of the confusion around this category. Surfer does not write copy. It tells you what the search engines want to see in your copy — which keywords to include, how long the piece should be, what questions your competitors are answering that you are not. This is research. It is the digital equivalent of reading the market before you write the brief, and in that capacity, it is excellent.
The real-time content scoring is particularly useful: you write, and Surfer tells you, as you write, how well your piece is likely to perform against the current competition. This is not a creative judgement. It is a structural one. And structure, despite what the romantics in our industry like to believe, matters. You cannot persuade someone who never finds your work in the first place.
Where Surfer fails — and this is not really Surfer's fault — is when people mistake its recommendations for creative direction. The tool will tell you to include a keyword seven times. It will not tell you how to include it without sounding like a parrot. That remains your problem, and it is the problem that separates the craftsman from the technician.
5. Claude
Claude ($20/month) is the most literate of the machines. I say this having tested every major language model over the past year, feeding them the same briefs, the same prompts, the same deliberately ambiguous instructions that I would give a junior copywriter on their first day. Claude's responses are, consistently, the most coherent, the most nuanced, and — this is the word I keep returning to — the most considered. It reads as if it has actually understood the question, rather than merely pattern-matched against it.
I have used Claude for research, for brainstorming (a word I dislike but cannot avoid), for testing arguments before I commit them to an essay, and for drafting client correspondence when I am too irritated to be diplomatic on my own. In each case, it has been useful in the way that a well-read colleague is useful — not as a replacement for your own thinking, but as a surface against which to sharpen it. It handles large volumes of text with a contextual understanding that its competitors lack, and it is capable of following complex, multi-part instructions without losing the thread.
It is not a copywriter. But it is, perhaps, the first machine I have encountered that I would not be embarrassed to have in the room during a strategy meeting. That is a small compliment, but from me, it is not a small one.
6. Descript
Descript ($24–65/month) has done something that I thought impossible: it has made video and audio editing accessible to people who are not editors. I include myself in this category. I have spent my career working with words on a page, and the transition to video content — which the market now demands, whether I like it or not — has been a source of considerable anxiety. Descript removes most of that anxiety by allowing you to edit video the way you would edit a document: you delete a word from the transcript, and the corresponding moment disappears from the footage.
This is, I think, genuinely clever. It respects the user's existing competence (in this case, the ability to edit text) and extends it into a new domain. More tools should work this way. The audio cleanup features are similarly impressive — removing background noise, adjusting levels, correcting the amateur-hour audio that characterises most marketing video content.
I do not pretend that Descript replaces a professional editor for work that matters. But for the daily production of social content, podcast episodes, and internal presentations, it is a liberation.
7. Adobe Firefly
Adobe Firefly ($60+/month) benefits from something that most AI image generators lack: institutional memory. Adobe has spent four decades building tools for visual professionals, and that history is evident in Firefly's interface, its integration with the Creative Cloud ecosystem, and — most importantly — its approach to intellectual property. Firefly is trained on licensed content, which means you can use its output commercially without the legal ambiguity that plagues its competitors.
The quality of the generated images is good, though not yet at the level where I would use them in a premium campaign without significant human refinement. Where Firefly excels is in ideation — generating visual concepts quickly enough that a creative team can explore ten directions in the time it once took to explore two. This is valuable. The early stages of a campaign are the most expensive, and anything that compresses the exploration phase without compromising its breadth is worth the investment.
It respects the Adobe tradition, which is to say it is powerful, somewhat overcomplicated, and essential.
8. Copy.ai
Copy.ai ($49/month) is adequate for what it is, and I choose that phrase deliberately. It generates emails, social media posts, content briefs, and product descriptions with a speed and consistency that makes it useful for teams producing high volumes of functional copy. Its brand voice feature maintains a reasonable consistency across outputs, and the unlimited word count on the Pro plan removes the anxiety of rationing that plagues some of its competitors.
What it lacks — and what, I suspect, it will always lack — is the ability to surprise. The copy it produces is correct. It is on-brief. It is, in the most damning sense of the word, professional. But it never makes you stop and think. It never produces a line that you wish you had written yourself. For production copy, this is acceptable. For anything that needs to move a human being from indifference to desire, you will need to look elsewhere.
9. Hootsuite
Hootsuite ($99/month) is not, strictly speaking, an AI tool. It is a scheduling and management platform that has incorporated AI features in the way that most legacy software has — sensibly, incrementally, without pretending to reinvent itself. I include it here because the discipline of scheduling is, in my view, one of the most undervalued aspects of modern marketing, and Hootsuite enforces that discipline more effectively than any tool I have tested.
The ability to plan a month of social content in advance, to see it laid out in a calendar, to ensure that messaging is consistent and properly sequenced — this is not glamorous, but it is essential. The AI-assisted caption generation is a minor convenience. The analytics are useful without being overwhelming. The real value is in the structure it imposes on a process that, left to its own devices, tends toward chaos.
Scheduling is discipline. Discipline is craft. I will not apologise for including a scheduling tool on a list of AI marketing instruments.
10. ActiveCampaign
ActiveCampaign ($49/month) is email automation done properly. I have a complicated relationship with email marketing. At its best, it is the most intimate form of commercial communication — a message that arrives in someone's personal space and asks, with whatever grace you can muster, for a moment of their attention. At its worst, it is spam with better typography. ActiveCampaign sits closer to the former than the latter, and for that it has my grudging approval.
The automation capabilities are sophisticated without being incomprehensible — a balance that its competitors frequently fail to achieve. You can build customer journeys that respond to behaviour with a specificity that would have seemed like science fiction twenty years ago: a prospect opens an email, clicks a particular link, visits a particular page, and the system responds with a message tailored to that exact sequence of actions. The predictive sending feature, which determines the optimal time to deliver each email based on the recipient's past behaviour, is the kind of intelligent automation that I find genuinely useful.
The integrated CRM is competent, the interface is (mostly) logical, and the learning curve, while steep, is not vertical. For businesses that take email seriously — and if you do not take email seriously, you are leaving money on the table — ActiveCampaign is the tool I would recommend.
A Final Principle
I have now spent several thousand words on a subject that I approached with reluctance and conclude with ambivalence. These tools are real. They are useful. Some of them are, in their narrow domains, genuinely impressive. But they are tools, in the same way that a typewriter is a tool, or a printing press, or a camera. They do not contain ideas. They do not understand what makes a human being reach for their wallet, or change their mind, or remember a brand twenty years after the campaign has ended. They process. They optimise. They automate. These are valuable functions, but they are not the function that matters most.
The function that matters most is thinking. Clear, disciplined, occasionally uncomfortable thinking about what you are trying to say, to whom, and why they should care. No tool can do this for you. No tool ever will. The practitioners who thrive in the age of AI will be those who use these instruments to eliminate the mechanical drudgery of their work, freeing themselves to spend more time on the part that no machine can replicate: the application of human judgement to human problems.
That is not a prediction. It is a principle. And principles, unlike tools, do not require a monthly subscription.