Agentic AI shakes up ad buying, but humans still rule

By early 2026, agentic AI systems like PubMatic’s AgenticOS and Optable’s Audience Agent demonstrated a breakthrough in privacy-first programmatic workflows, enabling real-time activation of rich first-party audience data without exposing raw information. Their pioneering use of the Advertising Context Protocol (AdCP) facilitates agent-to-agent collaboration that preserves privacy while avoiding closed platforms, marking a significant advance in how programmatic advertising can leverage audience intelligence securely and efficiently.

The Trade Desk’s Koa Agents, powered by Anthropic’s Claude, exemplify how agentic AI can automate complex campaign workflows—from media plan ingestion to troubleshooting—while maintaining human oversight and explainability. CEO Jeff Green highlights the challenge of managing thousands of variables in scaling campaigns, noting that Koa Agents help marketers navigate these trade-offs by framing decisions rather than autonomously executing buys, thereby preserving marketer control and trust.

Emphasizing data portability and advertiser empowerment, The Trade Desk’s Koa Agents return campaign performance data to clients, enabling them to train proprietary models and treat first-party data as a competitive asset within an agentic AI ecosystem. This approach aligns with a strategic shift away from agency intermediaries toward direct advertiser relationships, evidenced by a 55% year-over-year increase in joint business deals in March 2026, signaling growing adoption of AI-driven simplification in programmatic buying.

The August 2026 Kokai Zuma update further advances agentic AI workflow automation by introducing the Koa Assistant, a conversational AI interface that orchestrates multiple AI agents, including audience creation tools that suggest targeting strategies based on campaign prompts. These opt-in features prioritize explainability and human control, allowing integration with third-party AI via the Open Agentic Kit while enhancing transparency through streamlined reporting and real-time campaign insights, thus reinforcing advertiser confidence and operational efficiency.

By early 2026, the partnership between Optable and PubMatic marked a significant leap in privacy-first data collaboration by integrating Optable’s Audience Agent with PubMatic’s AgenticOS. This integration allows advertisers to activate and analyze rich first-party publisher data in real time without exposing raw data, thereby enhancing targeted campaign efficiency while upholding stringent privacy controls. This seamless collaboration empowers publishers through PubMatic Activate to improve addressability and yield, responding swiftly to advertiser demand without requiring changes to existing protocols.

Central to this innovation is the use of the Advertising Context Protocol (AdCP), which facilitates agent-to-agent workflows that enable programmatic advertising to harness privacy-safe audience intelligence beyond closed platforms. This pioneering approach by Optable and PubMatic demonstrates how decentralized, privacy-conscious data exchanges can power real-time audience activation, setting a new standard for privacy-first automation in programmatic advertising ecosystems.

CTM’s AskCTM AI agent, launched in mid-2026, exemplifies how conversational AI can democratize programmatic advertising by enabling non-technical marketers and agencies to self-serve campaign setup and configuration without engineering support. By guiding users through deploying AI-powered features like AskAI and VoiceAI, AskCTM not only accelerates adoption but also empowers those closest to customers to directly manage campaigns, bridging the gap between technical complexity and marketing agility.

The integration of OpenAI Ads attribution within CTM’s platform closes critical attribution gaps for ChatGPT-driven campaigns by linking ad interactions directly to downstream conversions such as calls, form fills, and appointments. This closed-loop tracking capability, a first in the industry, enhances marketers’ ability to optimize AI-driven campaigns with real-time conversion data, particularly benefiting verticals reliant on phone calls where traditional attribution has been challenging.

InMobi’s Buyer Hub introduces a hybrid conversational AI assistant embedded within its buyer portal that streamlines inventory discovery, deal building, and campaign optimization while preserving trader control through role-based access and curated inventory. This approach reflects a cautious but strategic embrace of AI assistance in programmatic buying, balancing automation with human oversight amid market data showing conventional buyers currently outperform AI agents on cost efficiency.

The Trade Desk’s Kokai DSP dashboard update, featuring the Koa Assistant, marks a significant step toward user-friendly automation by consolidating AI-powered tools for audience creation, ad frequency optimization, and performance insights into a single conversational interface. By integrating purchase-level data from retail media partners and surfacing key campaign metrics directly on the dashboard, The Trade Desk lowers technical barriers and enables marketers to make data-driven decisions more intuitively and efficiently.

By mid-2026, programmatic platforms like InMobi and The Trade Desk have embraced conversational and agentic AI assistants to streamline campaign workflows, enhancing efficiency without sacrificing human oversight. InMobi’s Buyer Hub integrates a conversational AI that aids inventory discovery and deal building while preserving trader control, distinguishing itself by embedding assistants within buyer portals rather than fully automating transactions. Similarly, The Trade Desk’s Kokai Zuma update introduces AI-powered tools such as the Koa Assistant to automate audience building and simplify campaign insights, reducing manual setup time and enabling advertisers to access real-time conversion lift studies directly on dashboards, thus accelerating performance evaluation and transparency.

Despite the promise of agentic AI, adoption remains cautious as conventional programmatic buyers currently outperform AI agents on cost efficiency, underscoring the market’s transitional phase where human governance is paramount. Platforms like PubMatic and MiQ demonstrate that agentic AI can deliver dramatic efficiency gains—cutting campaign setup times by up to 98%—yet these systems augment rather than replace existing real-time bidding infrastructure, maintaining human approval and transparency at the workflow’s core. This careful integration reflects a broader industry trend toward balancing automation with accountability to foster trust and control.

The emergence of open standards such as the Advertising Data Communication Protocol (ADCP) is pivotal in enhancing transparency and collaboration across the programmatic ecosystem. By standardizing communication and workflows, ADCP empowers buyers with greater visibility into fees and optimization opportunities, facilitating scalable agentic advertising collaborations. Early adopters leveraging these standards are gaining long-term competitive advantages by enabling faster, more transparent, and privacy-conscious media buying, signaling a shift toward AI-driven closed-loop attribution and automation that respects user privacy while improving market efficiency.