Sudipto Das, Vice President, Advertising Solutions APAC, PubMatic looks at how indie agencies are leading the way in agentic advertising...

While interest in agentic AI is high in APAC agencies, so is caution. There are questions being asked about the use of agents in advertising workflows: Where’s the proof? What does it actually improve? How do we know it won’t create complexity instead of removing it?

These are fair questions. In an economic climate where every dollar is scrutinised and media buyers are under pressure to show their impact on business outcomes, teams need to justify every decision, investment, and shift in strategy.

But buyer scepticism creates a catch-22. As with any new technological advancements, agentic advertising needs adoption to prove its value, but adoption depends on proof already being there. 

But while many are still waiting for that proof to emerge, the evidence is already here.

The proof is already on the table

It’s important to define what we mean by agentic, as there are still competing definitions across the industry. Agents are AI solutions that can independently execute decisions across complex workflows. They are capable of coordinating multiple functions in sequence rather than performing a single action in isolation.

For advertisers, this might mean an agentic system that can set up a campaign based on a brief delivered in natural language, monitor its performance over time, and automatically surface (and sometimes action) recommendations to steer it towards defined KPIs. Agentic advertising takes the reins for routine tasks, while the direction of travel continues to be driven by a human pilot.

This streamlining is especially valuable for independent agencies that have historically had few technical resources and capabilities. But agentic AI doesn’t just help democratise capabilities for independent agencies; it can also play directly to their natural strengths.

Independent agencies build their reputations by providing highly individualised services delivered by specialist teams. They offer agile teams with a high level of local knowledge and close client relationships, which work especially well in a market as fragmented as APAC. This dynamic can also create a willingness to test approaches within both agencies and their clients’ businesses. It’s an operating model, and a region, uniquely well-placed for putting agentic AI to work.

Today, cutting-edge independent agencies use agentic workflows to streamline campaign management and respond to performance signals while a campaign is live. Work that once required multiple touchpoints and manual adjustments is now supported by systems that analyse signals and recommend changes as campaigns evolve. 

Using PubMatic’s AgenticOS, launched in January 2026, agencies across multiple markets are using more than 20 agents to help run and manage everything from media activation to deal management and curation. These agents operate across the PubMatic ecosystem, continuously learning from more than 200 publishers and 100,000 sites to enable effective real-time buying without intermediaries. One fully-automated agent-to-agent campaign using AgenticOS saw an 87% reduction in campaign set-up time and CPMs landing at 15% lower than expected. 

Why agile teams are moving faster

Part of this success is a simple matter of size. Independent agencies have the benefit of regional agility. Having lightweight, localised teams (which themselves often possess a lot of operational independence) makes it easier to experiment and implement new technologies. Furthermore, testing on small, regionalised campaigns carries less complexity and risk than a multi-million-dollar global contract.

Additionally, many independent agencies choose to work with external partners that specialise in specific capabilities. That open-ecosystem approach allows agencies to adopt new solutions quickly, without committing to one partner and locking themselves into long development cycles.

This is useful when it comes to agentic AI, where the technology is developing at an incredible pace. Agencies that are comfortable collaborating with multiple external tech partners are finding it much easier to test, refine, and scale these capabilities.

What agentic AI is actually changing

There is still a tendency across the industry to think of AI in advertising as just another layer of basic automation that doesn’t yet drive any real business outcomes. In reality, it introduces a new approach to decision-making that allows teams to focus on the most impactful actions to address their KPIs.

Instead of relying on static rules or manual updates, agentic systems continuously review how campaigns are performing and automatically identify the opportunities available and actions to take. The user defines the parameters, and the agent operates within them to surface recommendations that marketing teams can review and approve.

In this way, routine work such as pulling reports becomes less time-consuming, and marketing teams gain time and space to focus on the strategy  that will shape a campaign or prioritise client relationships. User-friendly interfaces surface data in an easy-to-use and actionable manner. Ultimately, agentic systems allow advertising experts to actually do what they’re hired to do - pull together creative campaigns that meet strategic goals - without being bogged down in the nitty-gritty of the process.

The industry does not need to wait

What’s emerging is a clearer picture of what agentic AI can do in practice. Teams are finding they can refocus their energy to use their most valuable expertise, while staying firmly in control of decisions.

What happens next in the APAC digital ecosystem will depend on who chooses to be involved now and how they approach it. Those who take the time to understand how these systems work in real situations will have a stronger influence on how they evolve, will be able to guide the direction it is taken in next, and capitalise on their return faster.