Every SME owner and CEO I talk to has two specific questions about AI: how to cut costs, or how to grow revenue and margin. From what I've seen — confirmed by dozens of conversations with executives — these are the only two questions that matter in a small or mid-sized company. The rest of the public AI debate is irrelevant for an SME decision-maker.
Why "adding AI" neither cuts costs nor grows revenue
Most attempts to deploy AI in an SME look the same. The owner buys a tool, hands it to the team, and waits for results. A few months later it turns out costs haven't dropped, revenue hasn't grown, and the team is just as loaded as before — only with extra tools. From what I observe, the reason is always the same. AI doesn't add value to a role — AI breaks a role into tasks, some of which stop requiring a full-time person. Until the owner sees this and rebuilds the way work is divided, AI tools will not translate into either lower costs or higher revenue.
A quiet jurisdiction war — and why now is the right moment
What actually happens in an SME is a quiet war over jurisdiction — over who has the right to do a given task. A salesperson with AI starts writing offers that used to come from marketing. A customer-service specialist with AI analyses customer profitability that used to be the controller's job. An operator with AI documents a process that used to be outsourced to a consultant. Hence the key conclusion: now is the moment to settle this war consciously. First, map the processes that actually generate revenue and margin. Then reassign tasks within those processes — clearly stating who does what with AI, who controls the result, and which slices of work are simply no longer needed. In other words — set the jurisdiction yourself, before it sets itself.
Live measurement as the basis for cost and scaling decisions
Mapping and reassignment only make sense if the owner can see how the new arrangement works in real time. Hence the third, most often skipped step: wire measurement into every mapped process. Handling time, unit cost, output quality, number of human corrections after AI. Without it, a decision to cut costs or scale up is a leap in the dark. With it, it becomes an operational decision. The SME CEO no longer looks at quarterly summary results — they see live where the organisation is fast, where a human is the bottleneck, and where AI is generating errors. So if you're thinking about cutting SME costs or scaling with AI without measuring the processes, you're probably scaling chaos or cutting blind.
My conclusion for SME executives
Don't ask: "which AI tool should I deploy?" Ask instead: "have I mapped the processes that generate my revenue, are the tasks in them reassigned to people with AI, and do I see live how it works?" Because in an SME, the winner isn't the one who buys AI fastest. The winner is the one who fastest turns their company into a measured, controllable organism — where jurisdiction is clear, costs can be cut consciously, and scaling becomes a decision, not a hope.
Dear Reader. If you run an SME and want to discuss with me how to map your processes, reassign tasks with AI, and start measuring them live, get in touch.
Leszek Giza