AI is making marketing execution easier. That makes judgement more important.

Marketing has spent years trying to make execution faster.

Brief faster. Produce faster. Launch faster. Optimise faster.

AI has moved that forward dramatically. Work that once involved considerable production time can now be explored, developed and executed much more quickly.

That is genuinely useful.

It also creates a problem.

The easier marketing becomes to produce, the easier it becomes to produce more of the wrong marketing.

Execution was never the whole problem

When marketing isn’t delivering growth, it is tempting to start with the things that are easiest to change.

Adjust the targeting. Change the bids. Rewrite the ads. Produce more content. Test another channel.

Sometimes that is exactly what needs fixing.

But often the problem sits further upstream.

The positioning is weak. The proposition isn’t relevant enough. The creative isn’t doing its job. Media is being asked to compensate for a strategic problem. The audience is wrong. Or brand, demand, channels and sales activity are being managed as separate pieces rather than parts of the same commercial system.

Knowing how to optimise a channel is useful expertise.

Knowing whether the problem is actually in that channel is judgement.

There is a difference.

What senior judgement actually means

“Senior judgement” can sound like one of those phrases used to make experience seem slightly mysterious.

It shouldn’t.

In marketing, a lot of it comes from breadth.

Enough understanding of brand to recognise when consistency matters more than another creative variation.

Enough media knowledge to know when the problem is reach rather than conversion rate.

Enough creative experience to spot when the strategy may be sound but the execution simply isn’t landing.

Enough commercial understanding to question whether marketing is being asked to solve the right problem in the first place.

And enough knowledge across those areas to understand how changing one part affects the others.

None of that makes specialists less important. Specialists bring depth that a generalist cannot — and should not pretend to — replicate.

But growth problems do not necessarily respect specialist boundaries.

Someone still needs to diagnose across the whole system.

That is where breadth matters.

Modern marketing gives us an extraordinary ability to optimise individual components.

The danger is optimising the components while the system itself is wrong.

AI is particularly useful inside a strong framework

This is where I think some of the discussion about AI and marketing gets stuck on the wrong question.

AI can be extremely effective when it is working inside a clear framework.

A defined audience. Useful positioning. A relevant proposition. Good evidence. Appropriate constraints. A clear creative job. A defined role for the channel.

Within that environment it can help explore routes, challenge ideas, develop execution, create variations and move work forward at a speed that was previously difficult to achieve.

But somebody still has to determine what the framework should be.

That does not mean AI has no role in strategic thinking. It can be very useful there too: interrogating information, testing assumptions, exposing gaps, comparing options and helping develop an argument.

There is, however, a difference between using AI to contribute to a strategic decision and handing it responsibility for deciding what good looks like.

The quality of the work still depends heavily on the quality of the judgement directing it.

AI doesn’t just scale output. It scales decisions.

The familiar criticism of AI is that marketers can now make mediocre content more quickly.

I think that is probably the least interesting risk.

The bigger issue is that AI makes the consequences of decisions easier to scale.

Get the positioning wrong and that idea can spread across far more customer interactions.

Lose control of the brand and inconsistency can multiply quickly.

Choose the wrong creative direction and you can produce dozens of polished variations of something that fundamentally doesn’t work.

Misdiagnose the media problem and you can spend money very efficiently amplifying an argument buyers still don’t care about.

The output can look productive. Individual metrics may even improve.

But if the framework is wrong, faster execution does not solve the problem.

It gives you more of the wrong thing.

Capability matters more when execution gets easier

This is one reason I found Ipsos’s recent Marketing Anchors work, developed in collaboration with Mark Ritson, interesting.

The research surveyed 1,226 marketing practitioners across the UK, US, Canada and Australia. Only 35% met its benchmark for foundational marketing knowledge, while marketers with training were four times more likely to meet that benchmark.

The report’s own argument is that greater speed without sufficient grounding in marketing fundamentals increases the risk of poor decisions.

That research is principally about foundational knowledge and training. It does not establish my wider argument about AI or senior marketing leadership.

But the distinction it makes between speed and informed speed matters.

Technology can increase what a marketing team is capable of producing. It cannot, by itself, resolve whether the organisation is making the right marketing decisions.

That requires a combination of knowledge, evidence, commercial context and judgement.

And increasingly, it requires someone able to look across positioning, brand, creative, media, channels, data and execution and work out where the problem actually sits.

Not somebody standing above specialists dispensing opinions.

Someone able to use specialist expertise properly because they understand the job each part of the system needs to do.

The decisions before execution matter more

AI will continue to make a lot of marketing production faster and cheaper.

Good.

There should be less justification for slow processes, unnecessary production overhead and weeks spent making things that can sensibly be done in hours or days.

But reducing the cost of execution does not reduce the importance of the decisions behind it.

It increases their reach.

AI makes good judgement more scalable.

It does the same for bad judgement.

The advantage therefore isn’t simply being able to produce more marketing. Increasingly, that capability will be widely available.

The harder part is knowing what is worth producing, what problem it is supposed to solve, and what needs to be true for it to work.

That was always the job.

AI just makes the consequences bigger.