By Keith Arrowsmith, Marketing Director, Onetag
The iPod was unveiled 25 years ago, on 23 October 2001, with a simple proposition: take your music collection and put it in your pocket. It was not the first MP3 player, nor did digital music begin with Apple. MP3 files, CD ripping and other portable players were already changing how people consumed music. What the iPod did was make the experience easier.
Through seamless computer integration for finding, organising, transferring, and listening to music, the iPod brought together technologies and processes that had previously felt disconnected.
That distinction matters when looking at what AI could do to advertising. Agentic AI has the potential to remove layers of operational friction in the same way that the iPod made a complicated digital music ecosystem feel simple. More importantly, it could make an increasingly fragmented advertising ecosystem accessible through a unified operational experience, without requiring marketers to navigate every system individually.
But advertising has a complication that music files do not. Every decision affects someone else. A campaign influences the advertiser paying for it, the publisher carrying it, and the audience experiencing it. Simplifying the workflow cannot mean hiding the decisions.
Technology adapts to the user, not the other way around
Music has spent decades becoming more accessible.
The MP3 helped separate music from the physical format it had traditionally been attached to. The iPod and iTunes then made managing digital music far more intuitive. Streaming went further, making the underlying file even less visible to the listener. What mattered increasingly was access: choose a song, press play, and let the technology handle the mechanics.
The infrastructure underneath that experience was critical. Different formats, devices, software and services had to work together well enough for the experience to feel seamless.
Advertising is moving towards a similar question of access.
Programmatic already depends on systems communicating with one another at enormous speed. Yet marketers still encounter a fragmented landscape of platforms, interfaces, and data sources. Setting up a campaign, accessing inventory, diagnosing delivery issues or making changes can involve multiple tools and handoffs.
Agentic workflows could change that relationship. Instead of requiring a marketer to understand every interface, an agent can interpret an instruction, access the relevant systems and carry out defined tasks on their behalf.
The important shift is therefore reducing the amount of specialist knowledge required to make sophisticated technology useful.
The interface could disappear
The iPod succeeded partly because it made the underlying technology less important to the person using it. You did not need to think about how the music was being transferred. You needed to know what you wanted to listen to and access it easily.
Agentic advertising could move towards a similar model. A marketer could provide an objective and the relevant parameters, while connected agents handle more of the operational work behind it.
That could include creating or modifying campaigns, accessing live campaign context, identifying delivery problems, and taking approved actions to resolve them. Agent-to-agent workflows could eventually allow different systems to communicate directly rather than relying on people to move information between them.
For that to work, interoperability is essential.
The advertising industry has already seen the value of shared standards. OpenRTB gave different parts of the programmatic ecosystem a common way to communicate. Agentic systems face a related problem: they need a reliable way to discover tools, access information and interact with other systems.
This is where the Model Context Protocol becomes relevant. MCP provides a common connectivity layer between AI agents and external tools and data. Its value comes from making connections easier rather than requiring every system to build a completely separate integration.
Making complexity easier, not invisible
There is an important limit to the iPod comparison.
A music file does not have a commercial objective, a publisher relationship or a consumer experience to protect. An advertising decision does, and also requires humans to be in the loop for strategic decisions, judgement and accountability.
That means reducing operational complexity cannot mean turning the advertising ecosystem into a black box. If an agent is deciding where the budget goes, marketers need to understand what it can access, what it is allowed to change and what information is influencing its decisions.
This becomes even more important as agents move from recommendations towards action. Clear permissions can determine when an agent can act independently and when it needs approval. Guardrails can define the objectives, constraints, and level of autonomy within which it operates. Human oversight remains important for decisions that require judgement or fall outside those parameters.
Transparency also needs to extend to the signals informing those decisions. An agent can process vast amounts of information, but that does not make every signal useful. Context, placement quality, attention, creative engagement and performance data can all contribute to a more complete view of an impression. The value comes from understanding how those signals relate to the objective, rather than simply adding more data to the system.
What AI should leave behind
The iPod did not eliminate music. It changed the friction involved in accessing it. MP3s, streaming and the services that followed continued the same broader evolution towards making the format less important than the experience.
Advertising could be heading towards its own version of that shift.
The interfaces and manual processes that marketers currently use to navigate the ecosystem may become less important as agents take on more of the operational work. The complexity itself will not disappear. It will move into the infrastructure, standards and rules that allow those agents to operate safely across different systems.
That is why interoperability will matter as much as intelligence. If agentic advertising is going to become established, agents need to be able to plug into the ecosystem without every connection becoming a bespoke technical exercise. They also need to operate within standards that make their actions understandable and accountable.
The most useful AI in advertising may therefore be the technology that makes complexity feel simple without pretending that the complexity is gone.
The iPod changed what people expected from accessing music. Agentic AI could change what marketers expect from accessing advertising technology.
The real test will be whether the industry can make that experience easier while keeping the decisions visible, connected, and under control.
Originally Published on: LinkedIn
