AI Enables Creative ‘Magic,’ but Humans Still Make the Call, Says Digitas Chief
CANNES, France — The pressure on creative teams to automate everything is real, but the smarter play is using AI to eliminate the predictable so humans can chase the unexpected. That’s the view from one of the largest agency networks in the business.
Rather than treating generative AI as a replacement for creative talent, Digitas North America is deploying it as a “co-pilot” that handles the expected so strategists and creatives can focus on building emotional connections between consumers and brands. The approach includes synthetic audiences built on Epsilon data that can pressure-test creative concepts before they ever reach a focus group – or a real CMO.
“AI is a wonderful tool for our creatives to find the things that actually might be more expected, so that they can continue to go through the creative process to find the unexpected, because that’s what creates emotional connections between consumers and brands,” said Amy Lanzi, CEO of Digitas North America, in this video interview with Beet.TV at Cannes Lions 2026.
Synthetic audiences as creative sparring partners
The agency has built what it calls a Synthetic Audience Agent, which uses past campaign data and brand guidelines to predict how target consumers might react to creative ideas before production begins. The tool lets teams surface weak concepts early and refine stronger ones without waiting weeks for formal research panels.
“We used to do this when you would wait and you’d get to a focus group and try to find an insight,” Lanzi said. “Of course, it’s not real humans, but it’s a pretty good proxy. And so you can talk about different places to play. You can then look at the creative you’re making and feed the synthetic audience the creative to get them to react.”
The capability sits within a broader AI operating system Digitas calls Digitas AI, which also includes tools for building brand-safe agents and enabling dynamic personalized web experiences. Parent company Publicis Groupe recently agreed to acquire LiveRamp for approximately $2.2 billion, a deal framed as accelerating “data co-creation for smarter agents” by combining LiveRamp’s clean rooms with Epsilon’s identity graph.
Context is king, but now you can actually deliver on it
Marketers have talked about contextual relevance for years, but Lanzi argued the tools now exist to make it operational. The shift requires creatives to think beyond a single brand act and instead design content distribution systems that can react to context while remaining authentic.
“From a creative, I say to creatives, think about a brand and the content distribution system a brand should have so that they can have the right types of creative to connect with the core or growth audience with context,” she said. “You want a brand act that can react to context and still feel authentic at the same time.”
The stakes are rising. Emarketer forecasts US AI ad spending will more than double to $68.25 billion by 2030, with more than 80% of AI advertising in 2026 appearing adjacent to AI content – such as traditional search ads next to Google AI Overviews – rather than inside chatbot conversations. That means brands need creative systems designed for AI-mediated discovery environments, not just human-curated feeds.
Networked experiences and the influence economy
Digitas describes its creative philosophy around “networked experiences” – ideas that live across the full ecosystem where a consumer plays, from content placements to community conversations to their preferred LLM. Lanzi said the definition of a consumer’s network now includes influence, community chat, and AI interfaces.
“When we think about a creative platform, it’s really about building within the network that a consumer plays within,” she said. “That network is really about influence. And that network is not only where you are placing content, it’s also community, what’s happening in the chat, also what their favorite LLM is.”
A Forrester study released with the 4A’s found nine in 10 US marketing agencies now use generative AI, and half use agentic AI for marketing execution. But the research also warned agencies are leaning so heavily toward productivity gains that they risk commoditizing creative output – precisely the trap Lanzi said Digitas is trying to avoid.
CANNES, France — The pressure on creative teams to automate everything is real, but the smarter play is using AI to eliminate the predictable so humans can chase the unexpected. That’s the view from one of the largest agency networks in the business.
Rather than treating generative AI as a replacement for creative talent, Digitas North America is deploying it as a “co-pilot” that handles the expected so strategists and creatives can focus on building emotional connections between consumers and brands. The approach includes synthetic audiences built on Epsilon data that can pressure-test creative concepts before they ever reach a focus group – or a real CMO.
“AI is a wonderful tool for our creatives to find the things that actually might be more expected, so that they can continue to go through the creative process to find the unexpected, because that’s what creates emotional connections between consumers and brands,” said Amy Lanzi, CEO of Digitas North America, in this video interview with Beet.TV at Cannes Lions 2026.
Synthetic audiences as creative sparring partners
The agency has built what it calls a Synthetic Audience Agent, which uses past campaign data and brand guidelines to predict how target consumers might react to creative ideas before production begins. The tool lets teams surface weak concepts early and refine stronger ones without waiting weeks for formal research panels.
“We used to do this when you would wait and you’d get to a focus group and try to find an insight,” Lanzi said. “Of course, it’s not real humans, but it’s a pretty good proxy. And so you can talk about different places to play. You can then look at the creative you’re making and feed the synthetic audience the creative to get them to react.”
The capability sits within a broader AI operating system Digitas calls Digitas AI, which also includes tools for building brand-safe agents and enabling dynamic personalized web experiences. Parent company Publicis Groupe recently agreed to acquire LiveRamp for approximately $2.2 billion, a deal framed as accelerating “data co-creation for smarter agents” by combining LiveRamp’s clean rooms with Epsilon’s identity graph.
Context is king, but now you can actually deliver on it
Marketers have talked about contextual relevance for years, but Lanzi argued the tools now exist to make it operational. The shift requires creatives to think beyond a single brand act and instead design content distribution systems that can react to context while remaining authentic.
“From a creative, I say to creatives, think about a brand and the content distribution system a brand should have so that they can have the right types of creative to connect with the core or growth audience with context,” she said. “You want a brand act that can react to context and still feel authentic at the same time.”
The stakes are rising. Emarketer forecasts US AI ad spending will more than double to $68.25 billion by 2030, with more than 80% of AI advertising in 2026 appearing adjacent to AI content – such as traditional search ads next to Google AI Overviews – rather than inside chatbot conversations. That means brands need creative systems designed for AI-mediated discovery environments, not just human-curated feeds.
Networked experiences and the influence economy
Digitas describes its creative philosophy around “networked experiences” – ideas that live across the full ecosystem where a consumer plays, from content placements to community conversations to their preferred LLM. Lanzi said the definition of a consumer’s network now includes influence, community chat, and AI interfaces.
“When we think about a creative platform, it’s really about building within the network that a consumer plays within,” she said. “That network is really about influence. And that network is not only where you are placing content, it’s also community, what’s happening in the chat, also what their favorite LLM is.”
A Forrester study released with the 4A’s found nine in 10 US marketing agencies now use generative AI, and half use agentic AI for marketing execution. But the research also warned agencies are leaning so heavily toward productivity gains that they risk commoditizing creative output – precisely the trap Lanzi said Digitas is trying to avoid.
Series Videos
