Ali Ikhwan
By Kimeko McCoy | May 12, 2026
The advertising industry has reached a critical inflection point in its relationship with artificial intelligence. As the novelty of generative AI tools fades, the discourse has shifted from the theoretical potential of automation to the practical, often messy, reality of "agentic media buying"—the use of autonomous AI agents capable of executing tasks, making decisions, and managing workflows with minimal human intervention.
At the 2026 Digiday Programmatic Marketing Summit (DPMS), held in Palm Springs from May 6–8, the conversation surrounding AI was no longer about whether to adopt these technologies, but rather how much leash to give them. As brands grapple with the pressure to remain competitive in a volatile digital landscape, the industry is effectively splitting into two camps: the "autonomous enthusiasts" and the "human-in-the-loop" pragmatists.
The Core Debate: Autonomy vs. Oversight
The current landscape of programmatic advertising is defined by a philosophical divide. On one side are the early adopters, eager to offload the heavy lifting of campaign management to AI agents. These organizations are testing the limits of machine-led bidding, creative optimization, and real-time audience targeting.
On the other side, a significant contingent of brand marketers remains skeptical, maintaining a firm boundary where AI serves as a consultative partner rather than an executor. These organizations argue that while AI can process data at scale, it lacks the cultural intuition and strategic nuance required to protect brand equity.
Duluth Trading Company, a brand known for its distinct, humorous, and rugged identity, finds itself positioned squarely in the middle of this spectrum. For Duluth, the integration of AI is not a binary choice between "all-in" or "none-at-all," but a calibrated approach that balances operational efficiency with rigorous brand stewardship.
Duluth Trading Company: A Case Study in Hybrid Strategy
Ellie Uberto, director of marketing at Duluth, articulated the brand’s philosophy during a live recording of the Digiday Podcast at the summit. For Uberto, AI is a tool designed to remove the "grind" from the creative and analytical process.
"AI can get you to the finish line, and you get to spend all your energy crossing the finish line," Uberto explained. This sentiment captures the essence of the hybrid model: the AI handles the grueling, data-intensive labor, while the human team focuses on the final, critical decision-making processes.
Operationalizing the Agentic Workflow
Duluth has made specific choices regarding which functions to automate and which to guard. The brand currently delegates "low-level" tasks—such as technical bidding, data normalization, and the iterative management of creative variations—to AI agents.
However, when it comes to the "soul" of the brand—its specific tone of voice, its trademark humor, and its long-term market positioning—the company keeps a tight grip. According to Uberto, these elements require human oversight that current agentic systems simply cannot replicate.
Furthermore, the brand relies on a trusted agency partner to manage these agentic bidding systems. "We’re comfortable with it because we know that our agency knows us very well," Uberto noted. By outsourcing the technical management of these agents, Duluth avoids the need to become experts in the underlying prompt engineering or model selection, focusing instead on the strategic outcomes the systems generate.
Industry Dissent: The Case for Human-Centricity
While Duluth sees efficiency in delegating execution, other industry leaders remain wary of the risks associated with ceding control. Glenniss Richards, senior director of digital media activation at Bayer, offered a starkly different perspective during the summit.
For Bayer, the stakes of media activation involve complex regulatory, medical, and consumer-trust considerations that preclude the use of fully autonomous agents. Richards emphasized that while AI is an indispensable tool for data digestion and speed, it falls short when it comes to the nuanced context required for healthcare-related advertising.
"It’s making us quicker, faster, certainly more agile, giving us data to digest, consume, and inform our media campaigns," Richards said. "But it’s not owning or controlling our campaigns."
This hesitation highlights a recurring theme in the programmatic industry: the "trust deficit." For many brands, the transparency of AI models remains an issue. If an agentic system makes a bad decision, can the brand trace that decision back to its source? Without that level of auditability, many major advertisers are unwilling to grant agents the "keys to the kingdom."
Chronology: The Evolution of AI in Programmatic
The transition to agentic buying didn’t happen overnight. It is the result of a multi-year evolution in the ad-tech stack:
- 2023–2024 (The Generative Phase): Marketers focused on using LLMs (Large Language Models) for ideation, drafting ad copy, and generating synthetic assets. AI was primarily an assistant for brainstorming.
- 2025 (The Integration Phase): Companies began embedding AI models directly into DSPs (Demand-Side Platforms) and SSPs (Supply-Side Platforms) to automate basic optimization tasks like budget allocation and frequency capping.
- 2026 (The Agentic Phase): We have entered the era of the "Autonomous Agent." These systems do not just answer prompts; they are assigned "goals" (e.g., "increase ROAS by 15% for the spring campaign") and take a series of iterative actions—bidding, pausing, and adjusting—to achieve those outcomes without human intervention.
Implications: The Changing Role of the Media Buyer
The rise of agentic media buying has profound implications for the future of the marketing workforce. As agents take over the "mechanical" aspects of media buying, the role of the human media buyer is being forced to evolve from an operator to a curator and strategist.
The Shift in Agency Compensation
The traditional agency billing model—often tied to hours worked—is under threat. If an AI agent can perform a task in seconds that previously took a human team days, the value of those "billable hours" diminishes. Agencies are now forced to rethink their pricing structures, likely moving toward value-based pricing or performance-based incentives, where they are rewarded for the outcomes their agentic systems produce rather than the time spent managing them.
Transparency and Governance
The push toward agentic buying has brought the issue of "black box" algorithms to the forefront. As brands move further away from the manual controls of their media buying, the demand for transparency from tech vendors has intensified. Marketers are no longer satisfied with "it works"; they are demanding to know how it works, what data is being used, and what guardrails are in place to prevent brand-safety failures.
Future Outlook: What Lies Ahead?
The consensus from the 2026 Programmatic Marketing Summit is clear: the genie is not going back into the bottle. Agentic systems are moving beyond simple tasks and into the core of campaign architecture.
However, the industry is entering a period of "cautious acceleration." Brands like Duluth Trading Company represent the most likely path forward for the majority of the industry—a middle ground where AI is treated as a highly capable but strictly supervised employee.
For the remainder of 2026, we can expect the following developments:
- Standardization of "Agentic Guardrails": Expect the emergence of industry standards for how AI agents interact with brand-safety protocols.
- Focus on Strategy over Execution: The media buyers of tomorrow will be judged on their ability to set the correct objectives for AI agents and interpret the results, rather than their proficiency in clicking buttons within a dashboard.
- Increased Scrutiny of Agency Partners: Brands will demand that their agencies provide clear oversight reports that prove the AI is aligned with brand values, not just performance metrics.
As the industry continues to experiment, one truth remains: AI is a powerful force multiplier, but for brands with a distinct identity and a reputation to protect, the "human touch" remains the final filter of quality. Whether a brand chooses to lean into the AI revolution or keep a safe distance, the ability to manage the interface between human strategy and machine execution will be the defining skill set of the next generation of marketers.
Rifan Muazin
Ammar Sabilarrohman
