I recently had a conference call with a global procurement team interested in how we at TrinityP3 approach benchmarking the performance of AI within agencies and its impact on fees. It’s a question we are hearing more and more from both marketers and agencies.
In the high-stakes world of advertising, the promise of a fully autonomous, end-to-end AI-driven agency, like the recently announced Eversana, sounds like the ultimate game-changer. It's a vision of a frictionless operation where generative AI creates stunning visuals, predictive models identify the perfect audience, and process automation executes campaigns with machine-like accuracy. This story, often promoted by enthusiastic tech advocates, depicts a future of extraordinary efficiency, speed, and cost savings.
But let’s pause for a moment. This vision is more of a compelling fiction than a present-day reality. While the components exist, from LLMs that can churn out copy to algorithms that optimise bids, the seamless, holistic integration required for an "agency in a box" remains a distant horizon. The critical issues and considerations are not merely technical; they are deeply human, ethical, and strategic in nature. The real conversation we need to be having isn't about if this will happen, but what needs to be addressed before it can.
The 'black box problem: Ethics and explainability
The primary and most significant challenge is the 'black box' problem. As AI systems, particularly complex machine learning models, become more advanced, their decision-making processes become less transparent. An algorithm might decide that a campaign should target a specific demographic in a particular postcode, but it cannot explain why this choice is considered the most effective.
This lack of explainability poses a serious issue in an industry that depends on strategy and accountability. When a campaign fails, a human strategist can usually explain why—such as misreading a cultural trend, a flawed creative idea, or poor budget distribution. But what do you say when the "brain" of the agency is just a series of calculations that cannot be understood? "The AI said so' isn’t a convincing or professional reply, and it definitely won't impress a high-paying client.
This opacity is closely connected to ethical concerns. AI models are trained on extensive datasets, and if that data is biased, the resulting campaigns will continue and amplify those biases. We've already seen cases where ad-serving algorithms unintentionally discriminate against certain genders, races, or socioeconomic groups. In a fully AI-driven agency, for instance, a model might favour showing job ads for higher-paying roles more often to men and lower-paying roles more often to women.
This isn't a mistake; it reflects the flawed data it was trained on. A human creative director or strategist serves as a vital ethical check, ensuring campaigns are not only effective but also fair and just. Without this oversight, an AI-driven agency risks legal and reputational damage, as well as societal harm.
The inadequacy of generative AI for true creativity
While Generative AI has achieved impressive advancements in creating text, images, and even videos, it fundamentally lacks the key quality that defines great advertising: genuine human creativity and emotional intelligence.
Generative AI is, at its core, a sophisticated pattern recogniser. It analyses billions of pieces of existing content to produce something that is statistically likely to be "good." It can generate a technically sound ad, but it cannot grasp the nuances of a brand's tone of voice, the cultural context behind a joke, or the visceral emotions that make a campaign truly resonate.
The greatest campaigns in history, from Apple’s "1984" to Nike’s "Just Do It", were not driven by data analysis. They stemmed from a human creative's flash of insight, a deep understanding of human psychology, and a willingness to take risks.
An AI cannot feel the tension of a looming deadline or the thrill of a breakthrough idea. It cannot have a tough conversation with a client or build a relationship based on trust and mutual respect. These are the intangible but invaluable elements that form the foundation of a successful agency. Generative AI is a powerful tool for creativity, but not a replacement for it.
Bottlenecks are human, not process
The traditional approval process is definitely a bottleneck, and a more efficient, AI-supported solution is not only feasible but already being implemented. However, the main issue isn't the process itself; it's the fundamental need for human judgment at critical moments.
An AI can verify an ad for brand compliance and legal disclaimers, but it can't tell a client that their strategic brief is flawed or that their budget isn't enough to meet their objectives. It can't challenge a poor idea or shape a vision for long-term brand growth.
The human-in-the-loop is vital not only for ethics but also for strategy. An account manager understands the client's business, their competitors, and their long-term goals in ways an algorithm cannot.
They are the ones who can interpret a campaign’s performance data—which the AI has carefully provided—and turn it into a compelling story for the C-suite. They are also able to spot a potential pivot or new opportunity that a predictive model, based on historical data, might overlook.
The true opportunity is augmented, not autonomous
The most realistic and promising way forward isn't a fully AI-driven agency, but an AI-augmented one. This approach leverages AI's strengths: data analysis, pattern recognition, and automating repetitive tasks. It enables human talent to focus on what they excel at: strategic thinking, creative problem-solving, and building client relationships.
This is where the virtuous cycle of speed, cost, and effectiveness really comes to life. The savings from process automation can be reinvested not into more technology, but into the people who operate it. We should be training the next generation of strategists and creatives to be fluent in AI tools, understanding how to leverage them for better insights and more compelling creative work. The aim is to build an agency that is faster and more efficient, allowing human experts to spend more time on big ideas and deep client relationships that deliver truly exceptional results.
Ultimately, a fully autonomous advertising agency is an attractive idea, but it overlooks the deeply human side of the business. Advertising relies on connecting with people, and for now, the most effective way to do that is through a mix of human insight and technological power.
The future of advertising isn't driven by machines; it’s powered by a supercharged, data-informed, ethically guided human creativity. And at every step, we must ensure we're actually making a genuine step forward.
Darren Woolley is CEO of TrinityP3 and the author of Campaign Asia-Pacific’s regular column, Woolley Marketing.
