In 2016, I wrote a short piece questioning the rush toward efficiency. At the time, the debate centred on digitisation. Retail had moved online, services were following, and the logic felt unassailable. Food arrived at the door. Legal advice appeared through portals. Newspapers shrank into screens. The assumption was clear, if it is digital, it must be better.

A decade later the assumption has shifted, but the flawed logic is still seen everywhere. We still believe that if it is automated, it must be better.

What has changed isthe scale. In 2016 we were digitising distribution, but a decade onwe are automating our own cognition! Generative AI drafts strategy documents (quite well), builds creative assets (very badly), analyses data sets (brilliantly), writes code (better than most) and optimises media in real time. The marginal cost of producing marketing output has collapsed. Where months and multiple agencies were once required, small teams can now do the work in days. The commercial case for efficiency is compelling, particularly in uncertain markets where boards respond quickly to stories of productivity gains and cost compression.

AI can deliver double digit efficiency improvements across marketing workflows. Production accelerates, testing cycles shorten, campaign variations have multiplied expanentially. On a spreadsheet for the finance bods the case should be clean and persuasive. But productivity is not performance and the effects are not being felt.

The risk is that AI makes organisations much busier and much less human. More bloody, report, dashboards, campaigns and content. More automated journeys tracked showing a rise in engagement and activity, with real time reporting improvements and ridiculous output multiplies. BUT (shouty capitals for effectiveness) impact does not automatically follow. Efficiency seduces MBA's and been counters because it is measurable. Cost per lead falls (whoop whoop), time to publish shrinks (boom), headcount shrinks (kaaching). These linear metrics are easy to present and understand. Effectiveness is slower, harder and less tidy because the work we do builds trust, strengthens memory, increases price tolerance and clarifies what your company stands for. Those outcomes (brand loyalty, cultural vibrancy, clarity of purpose, consistency of message) compound enterprise value, but resist simple measurement.

I believe that in complex B2B markets this distinction matters more today than ever. Buyers are not short of information they are short of clarity. The typical buying group now spans multiple stakeholders, each exposed to an unrelenting stream of commercial messaging. In that environment, more volume does not cut through complexity, it amplifies it. Automation encourages dispersion, constant production, relentless optimisation. The brands performing best tend to move in the opposite direction. They focus on fewer, stronger ideas. They simplify their propositions. They build distinctive signals that endure. They concentrate creative capital rather than dispersing it across infinite iterations.

AI can and should support that discipline. It can surface insights from customer data, identify friction in the journey, and generate competent first drafts at speed. It reduces waste and frees up time for higher order thinking. But what it can't do is decide what a business should stand for. It cannot sense when a message resonates with risk, ambition or belief. It recombines patterns from the past. Strategy demands judgement about the future.

The confusion in 2026 lies in mistaking generative capacity for generative thinking. Because AI can produce an infinite stream of campaigns, we assume it can produce big ideas. Scaling output is not the same as creating meaning. Clarity of message, purpose and value must precede efficiency. Strategy must precede tooling and narrative must precede scale. Otherwise we simply automate the wrong things faster.

When I asked in 2016 whether an online NHS would be more effective at patient care or simply cheaper to run, the question was never anti technology. It was about alignment. Does the tool enhance the core promise, or merely reduce cost per transaction? That question still defines this moment. Every AI initiative should be held to the same test. Does it strengthen the distinctive story of the business, or dilute it into generic optimisation? Does it build trust and preference, or just speed and throughput? If the chief benefit is time saved, then it is an efficiency lever. If it deepens relevance and belief, then it becomes strategic.

Technology now runs horizontally through every organisation, like finance or operations. It is infrastructure. It is indispensable. But it is not the defining pillar of value. In marketing, that pillar remains creativity, human judgement and the ability to simplify complexity into something customers can believe in.

A decade on, the tools have changed dramatically. The underlying question has not. Are we building something that merely costs less to operate, or something that is genuinely worth more? That distinction still separates efficient businesses from effective ones.