AI finds it hard to unlearn something, or to un-mutate. This is problematic, because if you’re not finding difference today, it will harm your brand tomorrow, writes UM’s Dan Chapman.
The corporate world has become obsessed with AI – and with good reason. Companies
that talk up their AI capabilities tend to perform better on the stock market, draw confidence in their future investment potential and are considered more “innovative”.

There is a prevailing belief that AI-led automation will drive efficiencies, which is absolutely the case and something that brands can pin their hopes on. But without deep human insight in the mix, this could come at the cost of marketing efficacy.

Without human influence, intervention and ingenuity, AI’s normative mechanic and the current industry pervasive black-and-white decision-making, think ‘brand vs performance’, we face an existential threat to our industry: a landscape of ‘bland brands’.

Brand-first

The truth is, the advertising world has been increasingly oversimplifying the way brands exist for some time, without realizing that we are rejecting nuance for bluntness. This concerning trend neglects the fundamental purpose of a brand, which is to create ‘difference’; to set a product or a group of products apart in the minds of current or potential customers.

“A brand is a constellation in the customer’s mind,” says Jonah Sachs, one of the world’s leading experts on creativity. “It’s not just a logo; it’s a map of meaning that guides their choices.”

Taking Sachs’s stellar analogy further, consider this: would it make sense to ask an astronomer to reduce the galaxy to three points on a PowerPoint slide? Such oversimplification would do a disservice to the galaxy’s profound complexity. Similarly, reducing a brand to simplistic terms undermines its rich, individual essence.

A race to the mean

We’ve spent time testing various AI platforms, and our research in this space has shown us that some appear to use ‘genetic algorithms’. These mutate by favouring successful tactics and eliminating unsuccessful ones, mirroring Darwinian evolutionary principles.

While this approach is efficient, it will only be truly effective if it is considered as part of a total ecosystem – and fundamentally against a set of current and future marketing and business objectives unique to the brand itself.
A brand’s success ultimately depends on how effectively it interacts with the world – and how it evolves based on its actions and its successful (and unsuccessful) behaviours. This takes in the competitive landscape, economic conditions, cultural trends, and other external – and often complex – factors that impact customer interactions and brand growth.

Human ingenuity (powered by tech) allows us to recognize the complex nature of a brand’s existence and how the brand interacts with the world, which is essential for brands to harness AI effectively and avoid optimizing towards blandness. It’s a combination that means there’s no excuse to shy away from complexity. And we should certainly never accept being told to ‘simplify’, which is something our industry has been guilty of in the past (and still is).

We should never dumb down what we do; it devalues our craft in advertising. Strong brands – like Apple, Barbie and Kit-Kat – are complex in the ways they interact with the world and have deep and nuanced connections with people.

Instead of simplifying, we should use AI to help us understand these complexities more fully – using the computational power we have at our fingertips to structure the data, find the anomalies, and then through human ingenuity interpret it with more accuracy than ever before. This will ultimately allow us to create more meaningful and effective advertising.

F(orget) the funnel

Adland is in urgent need of transformation, with AI being both the protagonist and the antagonist, as we transition from the digital era to the era of intelligence.

A sensible first step would be to discard outdated, simplistic marketing models and heuristics, starting with the marketing funnel. Developed in 1898, this is a product of a simpler time, and 126 years later we must now move towards a model where we understand how the funnel sits within a more complex set of structures is key, to understanding the complex beautiful nature of a brand.

If you’re being shown a basic funnel of ‘awareness,’ ‘consideration,’ and ‘purchase,’ ask yourself: is that heuristic really doing my brand justice? Does the funnel feel right for marketing solutions and platforms with vast computational power and intelligence?

Whether human or artificial, the essence of heightened intelligence is the ability to detect and understand nuance, subtlety and difference. And so, in this intelligence era, if you don’t know what differentiates your brand – what its variance is – then your marketing will be ineffective. Why? Because, as you apply AI against non-differentiating strategies, planning and tactics (where brands default to ‘best-in-class’ and ‘best practice’) then the AI doesn’t learn in agency of the brand, it learns in agency of the category and the platform.
AI finds it hard to unlearn something, or to un-mutate. This is problematic, because if you’re not finding difference today, it will harm your brand tomorrow. To ‘do AI’ correctly it’s critical that end-to-end marketing heuristics are employed between the most creative strategists, brilliant data scientists and AI engineers – and that we don’t fall into the worrying industry trend of adopting rudimentary marketing tactics and esoteric behaviours.

Think beyond black and white

Our advice to clients is to ensure that, in advance of any algorithms deciding what to keep and what to kill, brands must first understand what the very best and very worst outcomes are. This should be supported by a solid understanding of the various other possible outcomes that lie in between. There can be no more binary (black and white) decision-making based on what is assumed ‘good’ or ‘bad’.

Factoring in a brand’s rich pattern of independent and interdependent variables that combine uniquely to affect and drive a unique outcome will result in ‘good learning,’ and drive future marketing efficacy. This approach is a model that is fit for purpose in this era, as this approach can maps to large language models. Why is this so important today? Because, in this world, a positive outcome today has a compound impact on a brand’s set of outcomes tomorrow.

There is a famous quote, apocryphally attributed to Einstein, that states humanity’s greatest invention is ‘compound annual growth rate’ (CAGR). I would argue that today, we can comfortably say that it is now ‘compound annual learning rate’ (CALR).

CAGR requires strategic and tactical investment forecasting to be successful. We need to take the same approach with a brand’s CALR. Brands must understand what failure actually is, what can truly be learned from it, and how it can be applied to a brand overall in a learning agenda that has systemized multi-variable experimentation at the heart.

Promoting and understanding the value of failure and risk must not just be a poster on a wall or only as part of a set of company values. It must be institutionalized into everything a brand and its partner agency deliver. Failure is learning. It is critical for future success.

Stand against bland!

The primary advantage that the significant computational capabilities of AI has is to enhance human ingenuity. It can help us better comprehend and navigate the complexities and subtleties of our beautiful human messy world.

Because generative AI’s mechanic is to converge to the mean to operate to normative rules and outputs, the role of human strategy and planning in advertising has never been more critical. It’s up to us humans to reject ‘normative rules’, to not default to over-simplification that stifle brand uniqueness, leading to brand blandness.
Our focus every day should be on unravelling the singular and beautifully complex nature of brands and how they connect with people. Our new rallying cry is ‘Stand against bland!’ We must resist the urge to oversimplify and overlook the nuanced nature of a brand, as the normative mechanic of AI becomes the dominant force, we know it will be.

It’s time to celebrate variance, subtlety, diversity and the different features that make each brand unique. Because all a brand is, is difference!