By Digital Business Director, Mohini Lakhani
“Won’t AI take our jobs?” was the rallying cry of anyone who dared to glance at ChatGPT in late 2022, profoundly catalysing the AI boom as we’ve lived and breathed it for the past 3 or so years. From copywriters igniting their creativity to data analysts double-checking their calculations, the initial AI boom felt less like a technological revolution and more like a productivity companion.
From Automation to Autonomy
As AI continues to evolve at pace, we are entering a new era defined by agentic systems, moving beyond tools that simply assist, to those that can act independently – namely, ‘Agentic AI’. The shift from rule-based automation to more advanced AI has already transformed how digital operations function, but these systems have still largely remained dependent on human input.
Agentic AI offers the possibility of changing this dynamic, introducing autonomous, goal-driven systems capable of making decisions, mirroring human-like neural networks and cognitive reasoning, whilst acting with minimal manual input. As this capability continues to develop, the key question becomes clear: does agentic AI represent a genuine step change in digital operational efficiency, or is it simply the latest fad?
Agentic AI at Scale
With much digital planning and buying operations currently centred around live data, rapid decision cycles, and constant optimisation, all strategically orchestrated by human operators, oftentimes, this is across fragmented systems.
Agentic AI has the potential to go beyond this. Examples could include dynamically allocating budgets between channels, adjusting targeting, and iterating creative based on real-time performance signals. To bring this to life, an agent could theoretically identify declining performance in one channel, reallocate spend to higher-performing segments elsewhere, and simultaneously launch new creative variations to improve results, all without manual intervention.
In this context, the discipline of agentic AI shifts paid digital operations. From what was once a model of continuous human management to one now powered by automation, underpinned by adaptive algorithms and data intelligence, agentic AI offers opportunities to revolutionise practices as we know them.
The Proof Points
We are already seeing this development from the tech giants. Spotify’s AI DJ, launched in Feb 2023, partially demonstrates agentic behaviour. Whilst it curates music for you in real time, adapting this based on your listening history and uses voice to guide the experience, it operates almost as a walled garden, i.e. only in Spotify, not acting outside the app, and is constrained by its capabilities.
Taking this one step further, both eBay and Amazon have integrated full agents into purchase flows to provide personalised, seamless shopping experiences. Tesco also recently announced its partnership with Adobe to better anticipate customers’ shopping needs, putting personalised offers in front of them, aligned to need states.
Whilst such brands have taken significant steps to implement agentic solutions, we’re also seeing strong momentum in the buy side of advertising, with platforms like Yahoo DSP following suit. For example, their agent allows for automation of campaign planning, activation, and optimisation, acting, in their own words, as an ‘always-on media buying assistant’.
Scope3 has gone so far as to coin AdCP (Ad Context Protocol), a protocol governed by AgenticAdvertising.org that enables agents to seamlessly communicate with one another across the advertising ecosystem. Imagine a world where the ad-selling-to-buying process is powered end-to-end with sole autonomy? Along with diminishing operational friction, for example, in time-consuming tasks like ad trafficking, this also gives the potential to increase transparency within the supply chain. In fact, Scope3 estimates that by 2030, 20-40% of campaigns will be delivered agentically.
The possibilities do seem somewhat endless, which demonstrates the velocity of agentic solutions within advertising practices and the opportunity to shift everyday operations.
The Maturity Reality Check
With developments accelerating at an unprecedented pace, it’s fair to ask: how far is too far? While preserving the human touch remains important, the trajectory is clear: AI is becoming more embedded in how we work, decide, and create. Adoption will only continue to rise, not as a distant possibility, but as an operational reality.
However, one truth still remains. Whilst it is “too easy” to allow developing agents like Google and Meta to automate campaigns, generate ad creatives, etc., these agents do still need training; thus, humans must remain in the loop.
If agents were left to run as they were, this could, and would, pose a significant risk to brand reputation. This makes robust governance frameworks not just desirable, but essential. As systems become increasingly autonomous, particularly with the emergence of agentic AI, which is now making decisions and taking actions on behalf of humans, questions around accountability, data security, and oversight emerge. Meta, for example, has already encountered challenges, including sensitive data exposure issues linked to early agentic implementations, underscoring how quickly innovation can outpace safeguards.
Taking a measurement lens, there’s also risk that as agentic AI embeds, the focus will fall to short-termism, this being the data that they are being trained on and therefore are making decisions on, is leaving gaps vs the longer-term impacts of media.
The Road Ahead
It’s becoming quite clear that roles won’t disappear as agentic AI continues to scale. Rather, they will simply evolve. The real differentiator won’t be whether AI replaces digital operations, but how effectively planners and buyers learn to direct, collaborate with, and govern with agents, almost resulting in an ‘agentic enterprise’.
With agentic growth, traditional media planning and buying roles are shifting their focus to strategic elements such as higher-order decision-making. Ultimately, this is where value lies for agencies as we look to the road ahead.