It’s time for an uncomfortable truth about AI in marketing: most of what gets pitched to SMEs is useless.
I’m not talking about the underlying technology. I’m talking about the application. The endless parade of chatbot builders, social media schedulers with a dash of GPT-4, and AI-generated-everything tools are solving problems that most small and medium-sized businesses do not have.
Having worked directly with over 300 SMEs at Creative Marketing Group, I see a chasm between the Silicon Valley narrative and the on-the-ground reality in places like Birmingham, Manchester, and Glasgow. The venture-backed obsession with SaaS products that offer marginal gains — a 10% faster workflow here, a 5% better-written email there — misses the point entirely. These are what I call ‘nice-to-have’ efficiencies. They don’t fundamentally alter the trajectory of a business fighting for survival.
SMEs aren't asking for more tools. They are asking for more outcomes. They want more cash in the bank, more sales, more high-quality leads. They want the result, not a slightly more efficient way to fail at getting it themselves.
The dominant model for AI adoption has been software-as-a-service. A business pays a monthly fee — £29, £99, £299 — for access to a tool that promises to revolutionise some part of their marketing. The problem is that a fool with a tool is still a fool.
Giving a struggling B2B company in Leeds an AI content generator doesn't solve their core problem. Their core problem isn't a lack of blog posts. It’s a lack of a coherent strategy, a non-existent distribution plan, and a fundamental misunderstanding of their ideal customer profile. The AI tool, in this context, simply helps them produce mediocre, untargeted content faster than ever before. It accelerates the failure.
We saw this firsthand. Clients would come to us, churned and frustrated, having spent thousands on a dozen different AI tools. Their marketing was a Frankenstein's monster of disjointed subscriptions, each 'solving' a tiny piece of the puzzle while the big picture remained a mess. The real blockage isn't execution; it's strategy. And strategy is a thinking problem, not a doing problem.
My contrarian take is this: the future of AI in marketing isn't more SaaS tools, it's the death of them. The future is agentic AI. It's about autonomous systems that don't just help you do the work but do the work for you. It’s a shift from providing a better shovel to just delivering the hole.
What does this shift from tools to agents look like? It's the difference between a grammar checker and a system that writes, targets, and sends your emails. It's the difference between a keyword research tool and an agent that identifies a content gap, commissions an article, optimises it, publishes it, and builds internal links to it.
This is precisely why I started building my own agentic AI products. With Autoemails, for instance, the goal wasn’t to create another Mailchimp competitor with better subject line suggestions. The goal was to build a system that our agency could deploy to manage a client's entire email marketing function. It ingests their sales data, understands customer segments, and executes entire campaigns autonomously to drive repeat purchases. The client doesn't need to log in, design a template, or A/B test a CTA. They just see the revenue.
That is what SMEs actually want. They don't want to become better email marketers. They want the outcome of great email marketing: sales. They want to buy back their time and offload the cognitive burden of a discipline they have no interest in mastering.
Take a successful UK retailer like Gymshark. In its early days, its success wasn't built on having the most sophisticated marketing stack. It was built on a deep, intuitive understanding of its community and relentless, targeted execution. The value wasn't in the software they used; it was in the strategic application of effort. Agentic AI is the first technology that can replicate the strategic layer, not just the executional one.