Is the future agentic?: Understanding new standards for AI in digital advertising

The industry is abuzz with the possibility of a world of autonomous AI systems. Agentic was the key theme of our recent Techtonic 2026 event, but the future can only be agentic if it’s built on common standards. 

Agentic was a hot topic in Palm Desert at the IAB US’ Annual Leadership Summit last week, and the consensus from industry leaders was the need for interoperability: 

"Interoperability is essential to maintaining performance while unlocking the next phase of automation." 

Ryan McConville, Executive Vice President, Chief Product Officer, NBCUniversal 

"Agentic advertising only scales when intent, execution, and governance evolve together. IAB Tech Lab's roadmap shows how agentic execution can be integrated into existing infrastructure while preserving interoperability." 

Nishant Khatri, Executive Vice President of Product Management, PubMatic 

"Our ecosystem operates best when we embrace and extend transparency and interoperability standards. This roadmap brings those principles to service-to-service and agentic transactions." 

Ray Ghanbari, Chief Technology Officer, Index Exchange

Why this matters now 

Here’s the reality: AI agents bring a performance challenge to Real-Time Bidding (RTB) that we’ve simply never seen before in programmatic. Unlike basic API calls that add predictable latency, AI agents require complex reasoning across multiple services, with each network hop adding 50-100ms to the already tight 200-500ms bid window. 

Traditional approaches just don’t move at this speed. When you need to activate segments, evaluate deal floors, and apply bid shading through separate external agents, the cumulative latency exceeds the entire auction timeframe. What makes this unique is the AI agent's need to process complex models in real-time. Traditional RTB can handle external calls for simple functions, but AI agents perform sophisticated computations that cannot be outsourced to remote services without breaking the auction timeline. 

To resolve this, IAB Tech Lab has introduced two key standards designed to keep the programmatic foundations running across direct, deals-based, and open markets. ARTF (Agentic Real Time Framework) enables AI agents to run co-located within host platform infrastructure through containerisation, eliminating network latency whilst maintaining data privacy. Agentic Audiences (formerly User Context Protocol) complements this by enabling signal exchange at machine speed through vector embeddings rather than text-based protocols. Together, they create an environment where AI agents can operate within RTB timing constraints without compromising performance. 

What are these new standards? 

ARTF: The execution framework 

Think of ARTF as a universal plug-in for our industry. Instead of every ad tech company building custom connections to every platform, ARTF creates a universal container format. It's like adopting USB-C: a single standard that works everywhere. 

What ARTF does: 

• Enables AI agents to participate in ad auctions safely and efficiently 
• Provides a container-based architecture where agents run in host platform infrastructure 
• Defines seven specific "intents" or actions agents can take: activating segments, activating deals, suppressing deals, adjusting deal floors, adjusting deal margins, bid shading, and adding metrics 
• Reduces latency by up to 80% by eliminating external network hops 

Agentic Audiences: The communication language 

Agentic Audiences provides a universal language for AI systems to share insights about audiences. It's like sending a compact data fingerprint that captures essence instantly, instead of a detailed paragraph that takes time to read. 

What Agentic Audiences does: 

• Standardises how AI systems exchange information about audiences 
• Uses "embeddings" (mathematical vectors with 256-1024 dimensions) instead of slow text descriptions 
• Enables sub-100ms communication times 
• Supports three signal types: identity, contextual, and reinforcement 
• Provides privacy-preserving communication through compact representations 

ARTF provides the execution framework (the infrastructure), whilst Agentic Audiences provides the communication language (the format). Together they enable AI agents to collaborate in real-time.

What this means for UK members 

For publishers: Faster bid processing means more inventory can be evaluated in each auction window, potentially increasing yield. The container model also gives you control over which agents participate and what data they access. 

For agencies and brands: Composable AI agents enable more sophisticated audience strategies without custom integrations. You could combine an insights agent, a creative optimisation agent, and a bid shading agent, all working together through standard interfaces. 

For ad tech providers: Standardised containers reduce the "one-off integration" problem. Build once, deploy across multiple platforms using consistent APIs.

At a glance 

• The 'How': ARTF is the engine (container framework), Agentic Audiences is the fuel (signal language) 
• The Goal: Reduce latency by up to 80% and enable smarter, faster AI collaboration 
• The Approach: Extends existing standards like OpenRTB, AdCOM, VAST, GPP, and TCF 
• The Status: Early stage (post-comment review). Time to learn, not yet time to implement widely

Building on existing foundations 

These standards build on established programmatic infrastructure, extending protocols like OpenRTB, AdCOM, VAST, GPP, and TCF, rather than replacing them. This is evolution, not revolution. 

November 2025: ARTF v1.0 released for public comment; LiveRamp donated User Context Protocol to IAB Tech Lab 

January 2026: Comprehensive Agentic Roadmap unveiled; User Context Protocol rebranded to "Agentic Audiences"; CloudX donated Agentic Mobile specification 

March 2026: All of Tech Lab’s Agentic work now falls under unified name: AAMP. AAMP consists of three pillars: execution, protocols, and an agent registry. Agent Registry launches in Tools Portal 

Current status: Post-comment review with industry feedback being incorporated 

IAB Tech Lab is integrating these proven standards with modern protocols, including Model Context Protocol, Agent2Agent, and gRPC, to support secure, machine-speed execution and scalable coordination between independent systems. 

"These standards represent compressed industry knowledge refined through billions of transactions. The fastest and smartest way forward is to build on an existing shared foundation, not introduce multiple new standards that create fragmentation. The industry gets the best of both worlds: high-performance agentic execution combined with the interoperability, governance, and trust the market already relies on." 

Anthony Katsur, Chief Executive Officer, IAB Tech Lab 

What's happening in 2026 

IAB Tech Lab is backing these standards with active development. On 1 March 2026, Tech Lab will launch an Agent Registry through the Tools Portal, enabling companies to register their agents for industry-wide transparency. Throughout the year, Tech Lab will also deliver open-source reference implementations of buyer and seller agents alongside standardised agent profiles. 

Additionally, Tech Lab is developing Protocol Buffers and gRPC mappings for existing specifications, alongside trust, provenance, and measurement signals for transaction integrity. 

For members wanting to learn more, Tech Lab hosted a public webinar on 28 January 2026 titled "Reviewing the Agentic AI Standards Roadmap," and is launching monthly in-person Agentic AI Boot Camps starting 12 February at the IAB Ad Lab (4:00–6:00 p.m. ET, second Thursday of each month).

Early stage, but clear direction 

Where things stand: 

• ARTF v1.0: Specification complete, public comment period ended, now incorporating feedback 
• Agentic Audiences: Open source development on GitHub 
• Production adoption: Early stage, with limited production deployments but strong industry interest 

Where to participate: 

• Explore the ARTF specification on GitHub 
• Contribute to the Agentic Audiences repository 
• Join IAB Tech Lab's monthly Agentic AI Boot Camps (starting 12 February 2026) 
• Participate in working groups to help shape the standards 

Preparing for what's next 

The agentic shift is coming, but let’s be clear: these standards are still in their early days. Adoption will take time, and a considered approach is required. 

Your action plan: 

1. Educate yourself - Understand what agentic advertising means for your business. Monitor the IAB Tech Lab GitHub repositories and review the specifications when you have capacity. 
2. Participate - Join working groups to help shape the standards. Attend the monthly Agentic AI Boot Camps to deepen your understanding. 
3. Consider pilots - Small-scale testing in late 2026 can inform your 2027 strategy. Use the reference implementations when they become available. 
4. Stay engaged - IAB UK will keep you updated with more resources as these standards develop. 

These standards provide a foundation for the industry to evolve together through open, interoperable approaches rather than proprietary walled gardens. By building on shared standards, we ensure that as AI becomes more central to advertising, the industry remains competitive, innovative, and collaborative.

Learn more 

Contact IAB UK for more guidance on agentic advertising and what these standards mean for your business. 

For a technical deep-dive, explore the ARTF specification or the Agentic Audiences repository.