At this year’s Digiday Programmatic Marketing Summit in New Orleans, agentic AI drew curiosity, confusion, and plenty of healthy skepticism. Across panels, roundtables, and hallway discussions, one message stood out above the rest: AI is gaining momentum, but the industry still prioritizes transparency, performance, and practical tools that deliver measurable results.
Here are my key takeaways from the December 2025 summit and what they could mean for buyers navigating the evolving landscape.
1. Agentic AI Is Promising, but Still More Hype Than Help
The event opened with a few jokes about agentic AI, but discussions quickly turned more serious. Most attendees agreed that the industry has not yet settled on a clear definition of an AI agent. Concerns about hallucinations, particularly in creative generation, remain a significant obstacle. Not a single agency reported using agentic AI for media buying today, and a quick poll of the room confirmed this.
The sense among some attendees and speakers was that the industry is attempting to create “agentic” solutions before a real problem exists. The consensus was that AI works best when it is targeted, measurable, and grounded in real data rather than replacing systems that already perform effectively.
2. Practical AI Works When Technical Foundations Are Solid
Even if marketers wanted to deploy agentic AI for media buying, the infrastructure is not yet ready. Agents cannot access the bid stream directly, latency remains an issue, and real-time autonomous optimization is still out of reach. Conversations naturally shifted to areas where AI can add real, measurable value.
My colleagues Margaret Williamson and Natalie Opirhory presented a sandbox demonstration of Cognitiv’s deep learning capabilities for contextual campaigns. We can scan billions of URLs to identify meaningful content, prioritizing context over simple keywords. This allows campaigns to capture timely cultural and seasonal moments while delivering performance-driven results—demonstrating how practical, data-backed AI is already helping marketers make smarter, more confident decisions today.
3. Transparency, Performance, and Customization Remain Top Priorities
Transparency, control, and performance dominated conversations. Rising identity costs, hidden platform fees, and the complexity of supply path optimization (SPO), especially in CTV, remain key concerns. Buyers want AI that is explainable, predictable, and aligned to KPIs, supporting humans rather than replacing them.
Agencies are also scrutinizing DSP and SSP relationships more carefully, focusing on cost-effectiveness, efficiency, and maintaining quality. Automation tools are helping buyers maximize performance, while SPO efforts remain focused on control and maximizing quality impressions.
Final Thoughts
My biggest takeaway was that the industry is not hesitant about AI, it is selective about unproven applications. Marketers want automation and intelligence that are transparent, brand safe, and tied to measurable outcomes.
Agentic AI may play a larger role in the future, but today belongs to solutions that help buyers work more efficiently, spend effectively, and achieve tangible results.
If you want to learn more about Cognitiv’s sandbox presentation or discuss how we can help you optimize your media strategy, contact me via email or LinkedIn.
