The evolution of AI in marketing: From big data to on-demand creative

Artificial intelligence (AI) has evolved over many years, and generative AI has everyone in the marketing industry wondering where they should invest their time and money - and the degree of caution with which they should embrace this technology.

This article is intended as a reference point for marketing teams to gain a common understanding of what AI is and how marketing has evolved with AI for years so teams can collectively understand the opportunities in front of them - and consider how they want to prepare for an AI-powered future.

What is AI in marketing?

In its broadest definition, it might be easiest to think of AI as a computer that can mimic the human brain's ability to learn, identify patterns and make decisions— but with the capacity to process information on a significantly bigger scale and speed than any of our brains ever could.

As a marketer you might be familiar with seeing suggested times to schedule emails and social posts - that's down to AI's ability to identify a pattern of when people open emails or social apps to predict the time you will get maximum engagement for your content.

It’s being intelligent in a way that mimics the human brain.

Where did AI in marketing begin?

So, where did it all begin? The development of AI traces back to the 1940s but in the 1990s and 2000s, as the internet took off and we got access to new sources of data - its application became mainstream for marketers.

We saw the introduction of sophisticated AI tools, including web analytics, search engine optimisation, and email marketing.

Businesses could now use AI-powered algorithms to analyse the vast amount of data collected from online activities and make informed decisions about how to target customers with personalised content that was tailored to their specific interests and preferences.

The rise of big data and machine learning in marketing

In the 2010s and 2020s, as our lives and businesses became embedded with online technologies, the world’s data continued to grow exponentially. This data, also known as big data, is the raw material for machine learning algorithms – powerful AI tools that can automatically learn and improve from experience without being explicitly programmed.

As platforms like Netflix and Spotify built their services with these algorithms creating sophisticated recommendation engines and personalisation, it marked a period of changing consumer expectations towards personalised content and marketing standards to be more customer-centric.

In marketing, they initially gained mass adoption with CRM systems that could forecast sales, automate personalised emails, and make recommendations based on past purchasing behaviours.

Deep learning and advanced consumer insights

Deep learning, a more sophisticated strand of machine learning, laid the foundation for hyper-personalised marketing. This is where AI-powered chatbots and digital ads that make you wonder if your phone is listening to you, marked a shift in customer-brand interactions.

Sentiment analysis tools like Agorapulse and Brandwatch came on the market to sift through vast amounts of data from social media platforms, blogs, and forums, enabling marketers to gauge public sentiment towards their brand or products in real-time.

Brands began deploying deep learning to optimise ad targeting, resulting in a more contextualised, effective, and non-intrusive advertising experience. An example of this in search would be if a consumer frequently searches for sustainable products, the algorithm in the ad targeting can ensure that advertisements for eco-friendly items are more prominently displayed to them.

Similarly, social media platforms like Instagram use deep learning to refine their ad targeting, ensuring that users see advertisements that are most relevant to their interests and behaviours. This level of personalisation and precision in marketing is all AI at work.

The evolution towards generative AI

While machine learning analyses patterns within datasets to make predictions, generative AI goes one step further; creating everything from text to images, and video, informed by the same data but with the ability to create original content that mirrors what a human might create. I’ve used it to help me write this article and my writing is better because of it.

This ability to create comes from using an advanced deep learning AI framework known as a Large Language Model (LLM) that is trained on vast amounts of text from the internet.

Wordsmithing, image generation, and video content can all be created with a few lines of instructional text - which we have come to know as a prompt.

An example of image creation using generative AI can be seen in the visual below for a digital campaign for travel broker Click & Go, which was created in Midjourney, a platform where users can create customised and unique visuals.

“With the Click&Go campaign, AI helped with the visuals, but the real magic happened when human creativity came into play. We had to dig deep into the brand, sketch things out on paper, and see if it felt right first. AI was just the icing on the cake after a lot of human brainstorming and tweaking.” John Martin - Art Director

While the adoption of generative AI tools continues to pace forward, it is important to remember that human oversight is critical. Human experience and creativity played the bigger role in bringing this campaign to life - Midjourney and Adobe Photoshop were creative tools used in the making.

Let’s look at another example. I have been experimenting with campaign analysis using OpenAI's ChatGPT - which makes it possible to upload documents for analysis. While I’ve gotten good results uploading Meta campaign data - there are big questions for organisations to consider before using AI in this way:

• How much will you trust its output over your industry and business knowledge?
• What data are you comfortable uploading to platforms like OpenAI’s ChatGPT when we don’t fully understand the extent to which it’s secure?
• Do you want your data to be used to train future LLMs?
• How is it helping or hindering your team’s ability to learn?

The opportunities to transform how our industry works are vast and so are the ethical, privacy and security considerations.

Preparing for the AI-powered future

Marketers now face a dual mandate – one to harness the incredible power of generative AI and another to understand its place within the complex ecosystem of marketing.

We must address issues like data privacy, consumer trust, and the potential for algorithmic bias, while also ensuring our teams are equipped to use AI effectively.

Whether you're a pioneer or a cautious adopter, the importance of human oversight cannot be overstated. The outputs from generative AI tools are incredibly convincing - and often wrong. They are full of bias because algorithms are only as unbiased as the data they are trained on.

Their conscious stream of what is right and wrong relies on us to guide them and it has never been more important to exercise the critical thinking and curiosity skills that are uniquely human.

Each marketer and each brand will have its unique dance with this technology. Some may choose to use it as an initial creative spark that supports idea generation. Others may integrate it more deeply to leverage its full predictive and creative capabilities.

Educating yourself and your team on AI and its implications is the first step and marketers need to prepare themselves and their teams for a future of continual learning, adaptation, and innovation.

If your team is interested in getting ahead of generative AI and the human skills needed for the changing workplace, see here for more details.