AI is not reducing costs, it’s redistributing them: Laurent Thevenet, Publicis Groupe

As generative AI moves from tool to infrastructure, agencies are discovering that speed is not the real disruption. The real shift lies in how creativity is produced, governed and valued.

The advertising industry has always evolved alongside technology. From the early days of digital to the rise of programmatic systems, each wave has promised efficiency, scale and precision. Yet the current phase of generative AI feels fundamentally different. This is not just another layer of optimisation. It is a structural shift in how creativity itself is produced and managed.

What is emerging is not simply faster output or smarter targeting. It is a redefinition of the creative process. AI is expanding the volume of possibilities, but it is also reshaping how ideas are filtered, refined and deployed. The linear model of creative development is giving way to a more fluid and iterative system.

At the centre of this shift lies a tension that the industry is only beginning to articulate. AI is accelerating output, but it is also introducing new complexities around governance, cost structures and creative decision-making. The promise of efficiency is being counterbalanced by the reality of managing systems that are constantly evolving.

For global networks like Publicis Groupe, the response is not about chasing speed alone. It is about building frameworks that can support rapid innovation while maintaining control. That balance is increasingly becoming the real differentiator.

Speed is rising. So is the need for control

Laurent Thevenet, Head of Creative Technology APAC at Publicis Groupe, describes the current pace of change as exponential. But he is equally clear that speed cannot come at the cost of structure. “The exponential curve is real,” he says. “But it doesn’t change our position on how we should fundamentally be structured from a safety, security, IT perspective.”

This signals a broader shift in how agencies are approaching AI. The conversation has moved beyond experimentation. AI is now embedded across workflows and is becoming an operational layer rather than a standalone capability. “Everyone is using AI, there is no question mark,” Thevenet says. However, widespread adoption does not mean full automation. What is emerging instead is a layered system where manual prompting, semi-automated processes and custom-built tools coexist. Creativity is no longer just about outputs. It is increasingly about the systems that generate those outputs. “We are seeing individuals creating their own software,” he says. “They could create their own software in a day.”

This marks a significant shift. Creative professionals are no longer just users of technology. They are becoming builders. The tools themselves are becoming part of the creative process, reshaping how ideas are conceived and executed.

Efficiency is real. But so are new costs

The dominant narrative around AI in advertising has focused on efficiency. Faster production, reduced costs and greater scale. But that narrative is beginning to show cracks. “AI systems are not cheap at all,” Thevenet says. “The cost conversation is not discussed enough in the AI space.”

While AI can reduce timelines, it also introduces new layers of cost. Compute power, iterative workflows and dependency on multiple systems are reshaping the economics of production. The outcome is not necessarily cheaper work, but differently distributed costs. “There are some campaigns we have launched in the past which would be done probably much faster today,” he notes. Yet speed alone does not define efficiency. The nature of creative work itself is changing. “You curate more than create,” he says.

This shift from creation to curation is becoming one of the defining characteristics of AI-driven workflows. Instead of building from scratch, creatives are navigating a vast pool of generated outputs, selecting and refining the most relevant ones. At the same time, AI can sometimes extend the creative process, rather than compress it. “People are now spending so much time iterating,” he adds.

This creates a paradox. AI enables faster production, but also encourages deeper and more continuous iteration. The result is a redistribution of time rather than a simple reduction.

Personalisation at scale still needs human judgement

India represents one of the most complex environments for AI-driven creativity. Its diversity of languages, cultures and regional behaviours makes it both an opportunity and a challenge for hyper-personalisation. In theory, AI is well positioned to address this complexity. It is trained on vast datasets that include cultural and behavioural signals. But in practice, execution remains dependent on human oversight.

“We should never underestimate AI,” Thevenet says. “But this challenge is very human. I’m not sure how good AI can be when hyper-localised.” Even in Southeast Asian markets where AI-driven localisation has been tested, human validation remains essential. “It needs constant quality check, which is very human, still, largely.”

 This highlights a key industry reality. AI can support personalisation, but it cannot yet fully replicate cultural nuance. The gap between data-driven insight and lived human context continues to require intervention.

The rise of new creative spaces

While much of the industry continues to focus on established channels, a different shift is taking place beneath the surface. AI is not just enhancing existing platforms. It is creating entirely new ones. Thevenet points to emerging areas such as optimisation for large language models and what he describes as “vibe coding.” This refers to the ability to generate functional applications in real time based on specific needs.

“Anyone can create and release apps today. Will they become a new channel? We don't know yet,” he says. This signals a move away from fixed media environments toward dynamic, on-demand interactions. Instead of designing for predefined platforms, creatives may increasingly design for moments. “The app will be whatever you need at a specific time,” he explains. This has significant implications for brand communication. It suggests a future where experiences are generated in real time rather than distributed across static channels.

The risk of sameness is real

As AI tools become more widely accessible, a new challenge is emerging. Creative homogenisation. “The risk is the sameness, for sure,” Thevenet says.

With shared datasets and similar models, outputs can begin to converge. This creates a tension between scale and differentiation. He observes two distinct approaches among creatives. Some use AI to replicate existing formats, while others use it to explore entirely new forms of expression. “Creatives are using it to redo what they used to do, or they are using it to create a new type of content.”

The latter approach is where future differentiation is likely to emerge. Experimental forms such as AI-driven surrealism are pushing the boundaries of what is possible. “With AI, you can break physics,” he notes. This expands the creative canvas, but also requires a shift in mindset from optimisation to exploration.

The industry is not chasing low-quality scale

Despite the rise of AI-generated content at scale, there is little evidence that brands are deliberately moving toward low-quality output. “We are not seeing brands doing that,” Thevenet says. While low-effort content may generate short-term attention, it does not align with long-term brand-building strategies. “At scale, no, I would not suggest that."

This reinforces a key distinction. AI may enable scale, but quality and intent remain central to brand value.

Creativity is expanding. But not evenly

The debate around AI often centres on whether it will expand or standardise creativity. The answer appears to be both. “I think for me at large, it’s expanding creativity,” Thevenet says.

AI is opening new possibilities in craft, execution and experimentation. However, the core of creativity remains rooted in human thinking. “The best ideas are still shaped by people." This duality may define the next phase of the industry. AI is an enabler of possibility, but not a replacement for human insight.

The real shift is still unfolding

What emerges from this transition is a shift in how AI is understood. It is no longer just a tool. It is a system that requires governance, collaboration and continuous interpretation. Publicis Groupe APAC has responded by building internal intelligence networks that allow teams to track and interpret rapid technological change. “We are all paying attention to what’s going on and asking what it means strategically for us,” Thevenet says.

This constant process of sense-making may become the most critical capability in the AI era. The industry often frames disruption in terms of speed and efficiency. But the deeper transformation is structural. It is about how creativity is produced, how decisions are made and how value is defined. AI is not just making advertising faster. It is changing what it means to create. And in that shift, the advantage may lie not with those who move the fastest, but with those who understand the system most deeply.