Will AI ever be better than humans at predicting what humans want?

Inside the transformative world of today’s most powerful Artificial Intelligence applications

“AI excels at finding insights in data and making predictions with unprecedented accuracy,” says Daniel Hulme , Founder CEO, Satalia , and Chief AI Officer at WPP. “This empowers creativity and strategic decision-making in profound ways.”

The combination of humans and AI has produced a new kind of talent. Right now, AI is being used to enhance the speed and effectiveness of human decision-making. It’s also helping to generate novel ideas. Machine-driven insights are powering human creativity in real time, enabling brands to craft and adjust their messaging in agile ways that resonate deeply with what their audiences truly want.

Brand conversation, like any conversation, is a two-way street, and it begins by listening. “Audience planning and audience intelligence is a big part of our business. That involves learning who we are trying to hit, and what narratives are currently being told about you. That’s very, very important,” says Rahul Titus , Global Head of Influence, Ogilvy .

Monitoring consumer opinion isn’t new of course, but rapid changes to the modern media landscape are causing massive challenges to brands. For Allison Spray , EMEA Chief Data & Intelligence Officer, Burson , the issue is a material one: “Traditional listening tools aren’t necessarily able to pull in material from the audio and visual content that dominates today. We can get the metadata, but that misses a tremendous amount of the conversation.”

According to Spray, this is where AI can step in. “My background is in measurement. I have spent the last couple of years looking at how we deploy AI to the problem of measuring communication outcomes and impact. And it can help. For example, Burson’s Sonar can deploy AI and ML to effectively watch or listen to that content, and very quickly pinpoint brand mentions and sentiment.”

Sonar can monitor this sentiment as it takes shape across many different channels. As Spray points out, this can help brands tackle damaging narratives as they develop, in real time. “Sonar’s ability to ingest and analyse audio and video content means that we can see where a conversation is unfolding on social media before it breaks into the mainstream. That gives our clients the ability to prepare, and even to engage and address any concerns that might be percolating,” she says.

But it’s not just about understanding the external conversation, but also how brand messages will be received. “What we’re looking for are the right tools, technology and talent to answer the questions that our clients put to us. Those questions could be how to understand an audience, what is going to motivate them to engage with our content, what is going to get them to consider one of our clients’ products for purchase,” adds Spray.

However, the dynamics of modern media environments – like virality – are complicating the traditional methods of prediction. It’s a point stressed by Titus: “In the past, just by looking at the number of followers somebody had, or looking at how previous content had fared, it was possible to predict future performance. But in the world of TikTok and Instagram, followers and past performance don’t matter anymore, and the virality of content cannot be as easily predicted. There are too many factors that go into it. At times it seems completely random.”

For Hulme, it’s no longer just about the numbers, but about how we understand what’s of real value, something AI can help brands discover: “From faster, better, cheaper, to ‘Does this make me attractive, virtuous, nostalgic, etc.’, AI is beginning to understand those elements of value, and how to activate them. It can even anticipate what humans will want before they explicitly express it, by understanding their underlying preferences and values,” he says.

When it comes to tools, Burson’s Decipher is proving to be a game-changer, helping brands look at more nuanced metrics such as ‘believability’ (how likely people are to trust a message) and ‘virality’ (how likely they are to share it online).

According to Spray: “Our clients want us to move beyond vanity metrics like clips, impressions, and into fully managing a campaign. Decipher is a cognitive AI solution that does that, helping us predict how people will think or behave. The intersection of believability and virality gives us what we call a ‘potential for impact’ score.”

This knowledge can then be applied anywhere across a campaign, from press releases to specific activations and calls to action. “We liken it to having a focus group in your pocket. We’re using it on everything from positioning statements to social content, and we’re increasingly working with the rest of WPP to think about how we can test it in the advertising space as well,” adds Spray.

The ability of AI to produce enhanced statistical modelling like this is transforming the world of marketing and advertising. But it doesn’t stop there. WPP’s unique “brains” – powered by Satalia’s expertise – are fundamentally changing how we understand and engage with human behaviour.

Daniel Hulme explains: “Our brains are not just large language models. While LLMs are statistical representations of data, we’ve developed specialised ‘brains’ that are trained on the collective genius and processes of our creatives and behavioural scientists. These aren't just mimicking human output; they're learning the underlying principles of creativity and human decision-making. For example, our creativity brain has already demonstrated the ability to generate award-winning ideas, and our behavioural science brain allows us to understand human nuances at a deeper level.”

“This to me is what’s powerful about generative AI. It’s not about just creating content but being able to recreate how people perceive content. We can simulate those signals, and create better content, and better predict neurological activation. Will this lead to clicks and likes, comments and sales? We can now ask that,” adds Hulme.

With all that knowledge, AI can also help you respond in the moment. It’s now possible not just to react to crisis situations, but to proactively deliver messages and campaigns at speed and scale, across a plethora of communications channels.

As Titus points out: “On platforms like TikTok and Instagram, the speed at which you react to culture is incredibly important. Think of something like Dove’s ‘Turn your Back’ campaign, which won gold in the Cannes Grand Prix in social and influencer. The team received a brief that morning, and within 24 hours they’d gone through kick-off, creative strategy, scriptwriting, production, to live deployment in seven markets. This is the power of AI: it helps you operate in real-time and be more agile than ever before.”

Hulme concurs, pointing out that, in a world of increasingly complex distribution networks, AI helps content find the right audience at the right time. “How do you ensure you’re pushing that content across your channels and gain the biggest impact? If you have five pieces of content, and five channels, then there are 120 solutions. If you have 15 pieces of content, and 15 channels, there are a trillion channels. Sixty pieces of content, and 60 channels? There are more options than atoms...”

For advertisers and marketers, ensuring authenticity remains paramount. “Any technology that has a material impact on people's lives has been massively, massively regulated. And we need regulation here too,” says Hulme.

This is especially true when it comes to influencer content, which is key to the success of most modern marketing campaigns. Titus cites the AI Accountability Act, which is intended to help maintain transparency in this area. “People trust people more than they trust brands – that’s why influencer marketing works. Therefore, if you betray the trust you have with consumers, you break the whole model.”

He goes on: “Last year, when AI was relatively new, there was so much false information – from virtual influencers to fake content – that people didn't know what was real and what wasn’t. That’s why we launched the Accountability Act, which asks the industry to disclose when they use AI. It allows us to ensure that content stays honest and ensures that brands and influencers are being authentic with their consumers.”

For Titus, this is of vital importance, given the pace of change and the increasing sophistication of consumers. “Authenticity is the number one thing that makes the industry work. WPP have been custodians of the influencer economy from day one. These creators have grown up with us – we’ve supported their livelihoods, and they’ve supported our growth. I take this duty of care very seriously,” he says.

It all comes back to the question of how we use AI to enhance what humans are creating, rather than letting the machines do everything, and to identifying what each are best at. This more reciprocal approach is what’s driving innovation, while at the same time helping to maintain the trust and authenticity on which the industry relies.

As Titus says: “The number one fact in my industry is that influencers know what their audiences want better than any marketer I've ever worked with, because they've built their audience one by one. So, we’re not using AI to tell them what content to produce, but to enhance how content is delivered and positioned. AI is not the solution, but the enabler. It’s not the Beatles – but it can be a very good George Martin.”

👉 Explore more AI‑powered marketing tools and capabilities on WPP Open