Roman Vrublivskyi is the experienced CEO of Attekmi, a global ad tech company that provides white-label programmatic solutions.
AI has been shaping programmatic advertising for years, and now the industry is entering a new stage of the technology evolution—agentic AI. Unlike traditional models that perform narrowly defined tasks, agentic AI can plan, make decisions, execute multistep workflows and adapt its actions with minimal human intervention. Basically, the opportunities offered by agentic AI explain its quick market growth. In 2025, the global agentic AI market size was valued at nearly $7.3 billion. By 2034, this figure is projected to surpass $139 billion.
For AdTech, the potential is significant, but the industry still faces a range of challenges. Let’s review them and find out what can be done.
Data Quality Keeps Playing A Critical Role
Autonomous systems can identify patterns and optimize workflows far more quickly than humans; however, the efficiency of their decisions depends entirely on the quality of the data. In programmatic advertising, this issue is especially relevant because data comes from multiple platforms, including ad exchanges, SSPs and DSPs. Differences in reporting methodologies, attribution windows, latency or missing signals can all influence how an AI agent interprets performance. For instance, with incomplete data, agentic AI may optimize toward metrics that don’t really reflect campaign success.
Creating more sophisticated AI models isn’t the most effective way to solve this challenge—ensuring stronger data foundations is much more important. This includes establishing consistent measurement standards across platforms, consolidating fragmented reporting, validating incoming data and improving data freshness. Implementing continuous data quality monitoring is essential as well.
Transparency And Explainability Become Even More Important
Nowadays, programmatic advertisers and publishers are increasingly willing to understand the logic behind AI decisions. However, with agentic AI, everything is complicated. It evaluates multiple variables simultaneously, makes independent decisions and continuously adjusts its strategy. This flexibility creates opportunities but also makes decision-making hard to understand.
Agentic AI should be designed with explainability as a core capability. Every significant action should be “accompanied” by a clear explanation and the data signals that influenced that decision. By keeping the decision-making process transparent, AdTech companies can build trust, retain users more effectively and gain a robust competitive advantage. In turn, businesses that fail to meet the growing demand for explainability can quickly fall behind.
Maintaining Brand Safety And Compliance Is Essential
At 60%, brand safety and suitability head the list of top causes for concern with programmatic advertising. Publishers don’t want irrelevant or misleading ads to appear on their media sources. Similarly, advertisers strive to prevent their ads from appearing next to low-quality content. However, agentic AI may aim for short-term efficiency. For instance, what if it prioritizes higher-paying demand even though it conflicts with publisher preferences or regional privacy requirements?
The risk of such issues introduces the next challenge—the power of agentic AI doesn’t remove our responsibility for ensuring brand safety and regulatory compliance. It’s critical to define clear operational boundaries that agents cannot override. These may include approved demand sources, geographic limitations, privacy requirements and so on. Besides, as the programmatic ecosystem continues to evolve, compliance monitoring and regular policy updates become crucial.
Balancing Autonomy With Human Oversight
One of the biggest questions about agentic AI is not whether it can make decisions independently but which decisions it should actually be allowed to make. Programmatic advertising involves a wide range of tasks, from routine operational tasks to high-impact strategic decisions. Logically, businesses aim to streamline as many processes as possible; however, allowing AI to operate without sufficient oversight can introduce undesired risks. For instance, agentic AI can deliver good results in bid optimization and anomaly detection. At the same time, entrusting it to make decisions about partner relationships is a tactic you should avoid.
The solution to this challenge looks obvious. Low-risk, repetitive tasks can be fully automated, while higher-impact actions require human approval before execution. However, it’s also critical to define thresholds, so that AI acts autonomously only when it has sufficient certainty. Finding the right balance between autonomy and human oversight may take time, but it is crucial for effective AI performance.
Integrating Agentic AI Into AdTech Is Technically Challenging
Building a capable AI agent isn’t easy, but deploying it successfully within the programmatic ecosystem is much more complicated. Modern AdTech infrastructure includes multiple technologies and platforms, each with different operational constraints, reporting formats, etc. For an autonomous agent, any inconsistency between these systems may reduce the quality of AI-driven decisions. Additionally, AI has to operate without introducing latency.
Successful adoption requires modern, interoperable infrastructure, but this isn’t something that can be achieved promptly. While the industry is trying to address this challenge effectively, businesses willing to integrate agentic AI into their programmatic advertising processes should do this gradually. Beginning with internal operational workflows and then expanding to critical optimization tasks is the approach that allows you to validate performance before enabling greater autonomy.
Final Words
Agentic AI can transform programmatic advertising by automating increasingly complex workflows, improving operational efficiency and helping teams make faster, data-driven decisions. However, unlocking the benefits requires more than sophisticated AI models. High-quality data, transparent decision-making, robust governance, human oversight and modern technical infrastructure are critical to ensuring autonomous systems operate effectively. As the technology matures, the most successful AdTech companies will be those that use agentic AI to support human expertise, not to replace it.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
