AI will not solve everything on its own. The winners will be humans with real craft who know how to use it.
We are learning something quickly: AI alone will not solve all our problems. Two things are changing how we work, and both come down to the same idea: humans plus AI, and end-to-end agentic processes. Underneath both sits a harder truth. The technology is moving far faster than the people who have to use it.
AI is only as good as our data
At its core, AI is data plus probability. The models predict, and the quality of the prediction depends on the data underneath. That means we need good, clean data at scale. Today, that is a human problem.
- Data sits in silos across systems and teams.
- It has gaps and is often unstructured.
- Legacy organisations feel it most, but any large, multi-system business struggles to bring its data together cleanly.
- Small pockets of clean data exist, but they do not unlock the full impact of AI.
The unglamorous work of unifying and cleaning data is the real starting line.
Stop automating steps. Reinvent the process
Think of a sausage factory. Many steps. Historically, each step was automated with its own machine. That is how we have treated marketing too, one manual step at a time. What is changing with AI and agentic systems is the ability to look at a process end to end. The real question is where it starts and where it ends. If we have a clear view of our inputs, structured as above, and a clear view of our outcomes, then the technology, the data and the AI can do the work in between.
The trap is automating small, formerly manual pieces and calling it transformation. Real change means reinventing the whole process. That needs a vision of the entire journey, which means a clear strategy and a clear view of where you want to go.
The real challenge is people
As in every revolution, the challenge is not really data, technology, hardware or software. It is people’s evolution. Technology is evolving light years faster than human beings. Biologically, we are not keeping pace with what we are building around us.
The capabilities marketers need to use AI well and reinvent our processes
A very clear view of what we are trying to do
That means strategy and creativity.A good understanding of the technology
Along with the ecosystems around us, so we can harness it.Too often, both are missing. Why?
The years before Covid were very digital-focused. The years after were very e-commerce-focused. Both pulled us away from traditional craft in media and creativity, and towards promo- and sales-driven performance marketing and last-touch attribution. The result is a whole generation of marketers trained to be very practical: create the product visual with the promo, optimise the performance media, watch ROAS. Miss the big picture.
But marketing is art and science, rational and irrational, and it is strongly driven by emotion. We need a reset around clear marketing fundamentals, back to creativity and to brand building in an emotional sense.
Measurement has got lazy
There is a second problem hiding in plain sight. For years we have leaned on last-touch and multi-touch attribution. They give one view of outputs and outcomes, and only part of the picture. That creates a double jeopardy. We only see part of the picture, because we do not have the data on the rest. And we believe what the models tell us anyway. What we need is holistic measurement. It is hard. It means going back to upper-funnel metrics, to brand, to understanding consumers, and bringing it all together in one model. That is the nirvana of measurement.
Instead, we got lazy. We picked the measures where the data is clean, structured and easy to turn into a conclusion, and so we only picked part of the problem. It is the same pattern as the data challenge: what is clean and available wins over what matters.
When an LLM gives you one answer, brand decides
In a world of ultimate choice, an LLM hands you one answer. The brand plays a critical role in whether you say yes or no to it. Emotion in choosing brands and products will matter even more than it ever did. Algorithms can narrow the options, but people still decide how they feel about the answer. Yet years of last-touch attribution and performance media have weakened the muscle that builds that emotional preference.
Back to the fundamentals
The fundamentals of marketing matter more than ever:
- Understanding consumers and teasing out insight
- Making a genuine human connection
- Building brands at multiple levels
- Understanding distinctive brand assets that go beyond the logo
- Building coherence across channels, and how they work together
- Seeing how it all fits in an ecosystem that is constantly changing.
This is the call back in my catchphrase: marketing hasn’t changed, but marketers must.
How will the next generation learn the craft?
Here is the second people challenge, and it worries me more. We know learning comes through pain and challenge. The entry-level jobs where you had to struggle to learn will not always exist, because technology is replacing them. The last 10 years did not build the skills we need, and the next 10 will hand many of these tasks to technology. So how do we teach these operational marketing elements, in school and in the professional environment?
I think it also changes what learning means. If the work that forced us to struggle is handled by machines, we have to create that challenge on purpose, to keep building the brain’s plasticity. Scott Young’s Ultralearning offers a useful starting point.¹ Its first principle is metalearning: before you start, work out how to learn the subject. Why, what and how. Learners need to understand how they learn, not just what they learn. And the struggle has to be designed in, through practice on real problems, testing yourself and seeking sharp feedback.
I grew up before pure technology, when the internet was 50K dial-up, and I learned some of this the hard way. Later, the Ritson mini MBA re-established some of the foundations for me. Even experienced marketers need to go back to basics. It is a deliberate choice.
The new world needs stronger capabilities
Two major changes are reshaping marketing: humans plus AI, and end-to-end agentic processes. Think of Excel. It did not remove accountants. It created more of them. In the same way, this new world demands even stronger human capabilities, above all strategy and creativity, which will never come out of a probabilistic model. Whether you see yourself as a cyborg or a centaur², we need to build our capabilities and our craft.AI will lift the bottom and create a vanilla layer of competent, similar-looking work.
"The winners will be the people with true marketing capability, a stretching vision and real creativity, who also use the technology."
Lex Bradshaw-Zanger
They will jump ahead, stand out from the crowd and win tomorrow.
How are you building craft in your team while AI takes over the entry-level work?
Notes
- Scott H. Young, Ultralearning (Harper Business, 2019). Metalearning is the first of its nine principles.
- The terms come from a study of BCG consultants by researchers including Fabrizio Dell’Acqua and Ethan Mollick, “Navigating the Jagged Technological Frontier” (Harvard Business School working paper, 2023). Centaurs split the work, handing some tasks to the AI and keeping others. Cyborgs weave their work together with the AI’s, moving back and forth.
Written by Lex Bradshaw-Zanger, editing by Claude
