As Advertising Week New York approaches, one conversation that will definitely dominate the agenda is agentic AI. Agentic AI has moved from a buzzword to a strategic boardroom priority, and the industry is beginning to grapple with how it will reshape workflows, business models, and the very definition of creativity.
OpenAI’s launch of ChatGPT Instant Checkout yesterday exemplifies just how quickly agentic systems can redefine the competitive dynamics of a sector like commerce and advertising.
The new era of agentic advertising
The future of advertising belongs to the agents. This statement has a double meaning. On one level, it refers to the rise of autonomous AI agents – systems capable of executing tasks and making decisions across the advertising cycle without constant human input. These agents are becoming essential in how ads are created, targeted, and optimised. But equally, the future belongs to those who can master these agents – the forward-thinking brands, platforms, and the agencies that learn to deploy them with skill. Importantly, this requires understanding that agents don’t run in isolation, rather they rely on the models, infrastructure, and orchestration layers beneath them to deliver results at global scale.
Unprecedented capabilities of AI agents
Agentic AI is powerful because it combines efficiency, scale, and personalisation. Imagine an always-on engine running your media strategy: agents analysing vast datasets in real time, generating creative variations, reallocating budgets moment-to-moment, and hyper-targeting messages to micro-audiences – all at a pace no human team could match.
We’re already seeing proof of this. Independent studies have shown that AI-driven ad campaigns consistently outperform manually managed ones, sometimes delivering double-digit improvements in return on ad spend. It’s evidence that AI agents can create measurable value today.
Building the foundation for agentic success
Indeed capturing that value is not as simple as switching on a new tool. Workflows built for human teams must be reengineered for AI collaboration. Talent must be retrained, not only in technical skills like prompt design and data curation, but also in governance and oversight. And organisations must rethink the human-machine relationship: letting agents handle the repetitive execution while humans concentrate on strategy, creativity, and ethics.
Critically, all of this needs to be considered as part of a wider AI strategy in order to generate true competitive advantage. Agentic systems sit at the top of a modern AI stack. To unlock these capabilities, agents need infrastructure, orchestration, and high-quality models to operate effectively. Without those layers, even the most intelligent agent can’t deliver reliable results at scale.
A defining moment for agencies and marketers
So what does this mean for agencies and marketers navigating this shift? Advertising has been here before. The shift to digital, the rise of programmatic – each wave of disruption forced reinvention. Today’s inflection point is intelligent automation at the core of operations. The choice is: lead the transition or risk having it imposed by competitors and platforms.
This doesn’t mean agencies disappear. But it does mean their value must be redefined. Repetitive optimisation tasks that once justified big fees are ripe for automation. The successful agency of the future will be one that designs and orchestrates hybrid workflows integrating human creativity with agentic efficiency.
Forward-looking leaders are already asking: What unique value can we bring in a world of agents? How do we codify our brand expertise into systems that reflect our values? How do we measure success when some of the work is automated? Agencies and brands that address these questions now will shape the hybrid future of advertising.
Taking action today
So how do you start? A few practical steps:
- Audit your workflows: Identify repetitive, data-heavy tasks from media optimisation to creative versioning that can be automated.
- Run controlled experiments: Let agents manage parts of campaigns under supervision, so your teams learn where they excel and where human input is essential.
- Codify your expertise: Translate your unique playbooks whether audience insights or creative frameworks into structured guidelines and training data for AI systems.
- Upskill your teams: Train people to supervise, prompt, and direct agents, the way a manager coaches a junior colleague.
- Build AgentOps: Set rules for how agents operate, what they can decide, and how humans intervene. Include governance, ethics, and compliance from the start.
My advice would be don’t wait until everyone else has already made the leap. Learn now, build trust in your systems, and create a competitive edge.
The human + AI partnership
Leadership in this era does not mean handing everything to machines. The most effective strategy is hybrid – human creativity plus AI productivity. Agents can do what they excel at: crunching data, testing endless variations, optimising every detail. Humans must double down on what only they can do: shaping big ideas, telling stories that move people, exercising cultural judgment, and protecting brand values.
As one marketing leader observed after testing AI-generated campaigns, “AI doesn’t have the human touch. Use it for what it does best, so people can focus on what they do best.” That is the balance. Humans provide vision and guardrails; agents provide scale and precision. Over time, the line may blur, but the competitive edge will always belong to those who orchestrate the best partnership.
What to expect in the next two years
Over the next two years, we’ll see agentic AI embedded deeply in advertising workflows. Media buying and optimisation are already being transformed by algorithms that continuously reallocate budgets and audiences. Personalisation engines are scaling one-to-one marketing. Routine creative tasks like resizing, copy adaptation, and A/B testing will be handled by agents.
And as adoption grows, governance will mature. Industry bodies and regulators are already moving toward clearer standards on transparency, bias, and intellectual property. But governance can’t be an afterthought or external layer. The most effective approach embeds security, oversight, and ethical controls directly into AI architectures, making responsible operation intrinsic to how these systems function from day one. This will make organisations more confident in rolling out agentic AI broadly. Expect to see the first high-profile advertising campaigns managed largely by AI agents, with humans in supervisory roles.
We’re already seeing hints of that future. During the 2025 NBA Finals, Kalshi, a sports-betting marketplace, aired a surreal AI-generated ad created in just three days for $2,000. It ran on national TV during Game 3, showing how quickly AI tools can move from prompts to broadcast-ready campaigns. While rough around the edges, it was a proof of concept: full creative execution at unprecedented speed and cost.
Inevitably, there will also be missteps. Just as early social media blunders forced companies to write new policies, we’ll see AI agents go off-course. That will drive better guardrails and governance. The organisations preparing now will weather those storms best.
The five-year transformation
Looking further out, agentic AI will move from micro-tasks to macro-responsibilities. Agents could manage entire functions – think of an “AI Jack” optimising long-tail campaigns or an “AI Jill” acting as a 24/7 community manager. These systems will become near-colleagues, learning continuously and contributing to strategy execution.
Humans will remain central, but their focus will shift. Instead of spreadsheets and trafficking, they’ll guide narrative, strategy, and ethics. Creative concepts might be brainstormed with AI partners; media plans drafted by agents and refined by human judgment. Consumers will still crave authenticity and emotional resonance, meaning human creativity will be elevated, not eliminated.
Trust requires guardrails
All of this progress depends on trust. AI agents reflect the objectives and data they’re given. Without strong oversight, they can optimise for clicks at the expense of brand equity or make biased decisions. Guardrails – technical, ethical, and cultural – must be built into systems from the ground up.
Organisations must also prepare their people. Teams should see AI as a partner, not a threat. Training, communication, and success stories will build confidence. Done well, AI adoption can boost morale by removing drudgery and freeing people to focus on higher-value, more creative work.
Why the full stack matters
All of this agentic potential rests on a critical foundation. While agentic AI is reshaping workflows, its true potential only emerges when integrated into a full-stack architecture that has been designed and optimised from the ground up to scale and adapt to changing environments. As we’ve seen in real deployments, the best outcomes are coming from systems where humans, models, and tools work together in a secure and efficient manner. That’s the real transformation that we at Stelia help our partners to build.
Are you ready?
The future of advertising will be shaped by the agents – and by the humans and systems that guide them. We are moving toward a world where campaigns are optimised in real time by algorithms, and creative concepts are co-crafted with AI partners. It’s a future of abundance – more content, more targeting, more personalisation – and therefore an even greater need for strong ideas and narratives to tie it together.
Those who embrace this shift with vision will find agents amplifying their impact and sharpening their strategies. Those who resist risk being outpaced by faster, smarter dynamics.
Advertising has always been defined by its ability to adapt to new technologies – from radio to television, from the internet to mobile. Agentic AI is the next chapter.
Adapt, experiment, set principles, and lead.
The agents are ready. The question is: are we?
