Marketers erect guardrails as AI agents reshape programmatic ad buying
Agency execs fear autonomous agents could blow a quarter's budget in a weekend.
The rapid evolution of AI agents in programmatic advertising is forcing marketers to balance automation with control. At Digiday’s Programmatic Marketing Summit, execs admitted they aren’t ready to hand over the reins. Henry Webster, SVP director of analytics at KSM Media, noted well-founded fears that an autonomous agent could “blow a quarter’s worth of budget in a weekend” due to hallucinations or incorrect CPMs. To mitigate this, KSM created an internal agent called “the librarian” that acts as a gatekeeper for each client’s brand voice, ensuring other agents use correct acronyms and campaign names.
Bayer takes even stricter measures. Glenniss Richards, senior director of digital media activation, described spend caps to prevent bots from overusing legacy data partners at the expense of new testing. The brand also requires data anonymization and de-identification before any activation. Transparency remains a major hurdle—attendees noted that LLMs can reinterpret intent over time, breaking out of guardrails. The IAB Tech Lab recently launched the Programmatic Governance Council (including WPP, Disney, and The Trade Desk) to create workflows and auction transparency standards. Until then, as Richards put it, “I want a person overseeing the bot.”
- KSM Media’s 'librarian' agent enforces brand voice guardrails by validating client-specific terms.
- Bayer enforces spend caps and data anonymization rules to prevent AI agents from overusing legacy partners.
- IAB Tech Lab’s new Programmatic Governance Council aims to standardize transparency, with members including WPP, Disney, and Amazon Ads.
Why It Matters
Marketers are racing to implement AI agents while limiting risk—a delicate balance that will define programmatic advertising's future.