By Yuichi Toyoda, CEO, dentsu APAC; and Chris Bower, Managing Director, Business Transformation (BX) APAC, dentsu
Each year, the Cannes Lions Festival of Creativity captures the pulse of our industry, and this year, AI dominated the conversation. But unlike previous years, when the focus was on experimentation and hype, the 2025 discussions reflected a shift: AI has become a core business capability, reshaping how organisations bring ideas to market.
Interestingly, some of the Festival’s most celebrated work reminded us that great creativity still begins with human craft. AXA’s “Three Words”, which won the Titanium Grand Prix, was pure copywriting genius; L’Oréal’s “The Final Copy of Ilon Specht” honoured a human voice that shaped culture decades ago. Even when AI featured, as in Natura’s “Amazon Greenventory,” it served human ideas rather than replaced them.
This contrast points to a larger truth: we’re still learning how humans and AI should work together. In many organisations, the two are still competing for creative space, or at best, operating side-by-side. The real opportunity lies beyond that balance: when human imagination and AI’s computational power are orchestrated to create outcomes neither could achieve alone.
That’s what we mean by Human × AI. It’s not a plus sign, but a multiplier. Human intuition, vision, and judgement combined with AI’s scale, speed, and pattern recognition to expand what’s creatively and commercially possible.
Why aren’t we there yet?
You might expect that several years since the rise of Generative AI, businesses would be using these tools more and more for idea generation. But the opposite has happened. According to a 2025 report by Filtered.com, “idea generation” has actually declined as a top AI use case compared to 2024.
There are three key reasons why.
First, most AI models still learn from the same public datasets.
So when everyone asks the same models the same questions, the answers start to blend together. Outputs feel generic because the inputs are generic. Businesses will only see differentiated ideas when they begin training and prompting AI using proprietary datasets with their own customer insights, creator inputs, cultural signals, and contextual knowledge.
Second, creativity doesn’t follow a straight line.
It’s emotional, iterative, sometimes messy. It loops, it meanders, it contradicts itself before it clarifies. Traditional AI systems weren’t built for this kind of non-linear process. They were designed to optimise, not wander. And wandering is where imagination often happens.
And third, most tools focus on AI’s ability to respond, but not its ability to collaborate.
Creativity isn’t a single prompt–response moment. It’s the back-and-forth: questioning, challenging, reinterpreting, refining. This interaction flow is where breakthroughs emerge, but most systems still underserve this part of the process. They give you an answer; they don’t help you build the answer with them.
To unlock AI as a true multiplier of human creativity, three conditions must come together:
- Proprietary data
- Workflows that reflect real creative behaviour
- A Human × AI flow that enables seamless iteration
These aren’t AI problems, they’re design problems. And solving them is what will allow AI to finally elevate, not imitate, human imagination.
Where are we headed?
At dentsu, we’re fortunate to start with a rich foundation of data: consumer behaviour, cultural signals, social insights, and future trends. We combine this with our clients’ own data, whether that’s customer information, product usage, or even the inner workings of their technologies. And critically, we identify where the gaps are and use a variety of techniques to fill them, so we’re not creating ideas from partial inputs.
From there, exploration and ideation happen through a multi-step, iterative workflow. Human teams provide context and direction at each stage, and AI runs structured processes in between. All of this flows into AIQQQ, our proprietary Human × AI innovation system, which rapidly generates new ideas based on the combined intelligence of people, data, and models.
But idea generation is only the beginning. Each concept is then reviewed by dentsu and client teams, discussed, refined, and fed back into the system. AIQQQ reinterprets the feedback, strengthens the insights, and elevates the most promising ideas, transforming good concepts into great ones.
Two things make this system fundamentally different.
First, its ability to absorb and contextualise a wide range of inputs.
We’re not searching for ideas in a vacuum. We’re combining research, client data, cultural context, and expert judgement into a unified innovation flow.
Second, AIQQQ is designed to replicate dentsu’s distinctive creative thinking process.
We’ve intentionally encoded the way our teams probe, reframe, and shape ideas, so the system reflects not just our insights, but our method of generating them.
Turning good ideas into great ones requires widening the lens. That’s where dentsu’s deep consumer research becomes powerful, especially when paired with client data. Our AI personas are built with far more nuance than typical models. They have richer motivations, clearer needs, and more dimensional preferences. This allows us to hold natural-language conversations that reveal how well a concept resonates and where it needs to evolve.
This Human × AI framework allows us to continuously innovate the process of innovation itself. Humans have an extraordinary capacity for imagination, but today’s business environment asks us to work at a speed and scale that no individual or team can match alone. At the same time, AI on its own cannot meet the standard of human creativity or judgement.
Only when we combine these strengths, seamlessly, intentionally, and iteratively, do we unlock the next frontier of business transformation.
