Ad Context Protocol (AdCP) launches for advertising automation

The advertising technology sector gained a new open-source framework on October 15, 2025, designed to enable artificial intelligence agents to interact with multiple advertising platforms through a single standardized interface. Ad Context Protocol, built on Anthropic's Model Context Protocol, provides a unified communication layer for managing campaigns across different ad tech systems without requiring custom integration work for each platform.

The protocol addresses long-standing fragmentation issues. Each advertising platform currently maintains proprietary application programming interfaces with distinct workflow requirements, documentation standards, and reporting formats. This forces media buyers and agencies to develop separate integration capabilities for Google Ad Manager, connected television platforms, digital out-of-home networks, and other advertising channels.

Brian O'Kelley, who founded AppNexus in 2007 and served as chief executive officer until its acquisition by AT&T in 2018, leads the protocol development. The project launched with version 2.0.0 on October 15, 2025, introducing what its creators describe as production-ready capabilities for modern advertising workflows.

The Ad Context Protocol is supported by six founding members: PubMatic, Scope3, Swivel, Triton, Optable, and Yahoo. Each founding member represents a distinct segment of the advertising technology ecosystem. PubMatic operates as a supply-side platform connecting publishers with buyers. Scope3 provides carbon emissions measurement for digital advertising. Swivel delivers audience targeting and analytics capabilities. Triton specializes in audio advertising technology for streaming and podcasting. Optable focuses on privacy-first data collaboration. Yahoo brings scaled media properties and advertising infrastructure.

The list of launch members extends to 23 organizations beyond the founding group, including AccuWeather, Adgent, Bidcliq, Butler/Till, Classify, HYPD, Kargo, Kiln, LG Ad Solutions, Locala, Magnite, Media.net, MiQ, Nativo, Newton Research, OpenAds, Raptive, Samba TV, Scribd, The Product Counsel, and The Weather Company. This roster represents both supply-side and demand-side participants in programmatic advertising.

The protocol's architecture operates through three distinct phases: discovery, comparison, and activation. During discovery, users describe target audiences in natural language rather than navigating platform-specific targeting interfaces. The system searches connected platforms for matching inventory and returns results in a standardized format. Comparison presents pricing, reach, and targeting capabilities in consistent data structures. Activation launches campaigns across selected platforms through single commands while maintaining platform-specific optimizations.

According to the technical documentation, the Ad Context Protocol provides 9 core tasks covering the complete advertising lifecycle. The get_products function discovers advertising inventory using natural language campaign briefs. The create_media_buy function launches campaigns with budget allocation, timing parameters, and promoted offering specifications. The get_media_buy_delivery function tracks real-time performance metrics including impressions, spend, click-through rates, and platform-specific engagement measures.

The technical implementation uses JSON Schema validation for all data exchanges. Each task defines input parameters, response structures, and error conditions through machine-readable schemas that enable automated client code generation. This approach reduces integration complexity by providing consistent validation rules across all protocol operations.

Response times vary based on operation complexity. Format listings complete within approximately 1 second through database lookups. Product discovery requires around 60 seconds for inference and retrieval-augmented generation processing. Creative synchronization operations span minutes to days depending on asset processing requirements and approval workflows.

AdCP supports asynchronous operations as a fundamental design principle. Operations return one of three status values: completed for immediate results, working for processing expected within 120 seconds, and submitted for long-running operations requiring hours to days. Human-in-the-loop workflows enable manual approval requirements at publisher discretion.

Campaign creation demonstrates the asynchronous approach. Simple campaigns may return completed status immediately. Complex campaigns requiring validation checks return working status during setup processes. Campaigns requiring human approval return submitted status with estimated completion times. Orchestration systems must handle these pending states through polling mechanisms or webhook callbacks.

Creative management capabilities include the build_creative function for generating assets from campaign briefs, the preview_creative function for visual preview generation, and the list_creative_formats function for discovering format specifications. The protocol includes a standard formats library covering video, display, audio, digital out-of-home, and carousel creative types.

The protocol's pricing model distinguishes between fixed-price inventory and auction-based purchasing. Fixed-price products specify cost-per-thousand-impression rates and minimum spend requirements. Auction products accept bid prices from buyers competing in real-time auctions. Some products support both pricing models, allowing buyers to select their preferred purchasing method.

Targeting capabilities operate through structured dimensions covering geography, demographics, interests, behaviors, content context, and temporal parameters. Publishers expose available targeting through their product definitions. Buyers apply additional targeting overlays during campaign creation to refine audience parameters beyond product defaults.

Budget allocation follows a package-based model. Each media buy contains multiple packages representing different product selections, creative format assignments, targeting configurations, and budget allocations. The system calculates impression delivery based on package budgets and product pricing, automatically adjusting calculations when buyers modify spending levels.

The Ad Context Protocol provides both Model Context Protocol and Agent-to-Agent Protocol access methods. MCP integration enables direct tool calls from AI assistants like Claude. A2A support facilitates complex workflows requiring multi-agent collaboration and approval chains. Both access methods provide identical functionality through different transport mechanisms.

Platform implementation varies across advertising systems. Google Ad Manager creates Orders containing LineItems that map to protocol packages. Kevel creates Campaigns containing Flights corresponding to packages. Triton Digital follows similar Campaign and Flight structures optimized for audio advertising workflows.

Creative synchronization uses an account-level library approach where assets upload once and assign to multiple campaigns. This addresses industry requirements for creative reuse across different media buys while maintaining centralized asset management. The sync_creatives function supports bulk operations with upsert semantics, allowing updates to existing assets or creation of new entries in single operations.

Performance monitoring extends beyond basic metrics. AdCP tracks delivery pacing against campaign targets, monitors budget consumption rates, provides package-level performance breakdowns, and enables dimensional reporting across targeting parameters. These capabilities support data-driven optimization through budget reallocation, targeting refinement, and creative rotation decisions.

The protocol's development follows open-source principles under MIT licensing. The GitHub repository contains complete documentation, reference implementations, and validation test suites. Monthly working group meetings address protocol evolution, implementation challenges, and future feature development. Public format discovery operates without authentication requirements, enabling broad ecosystem participation.

Security considerations include authorization verification for publisher properties, prevention of unauthorized inventory resale, and protection against fraudulent campaign creation. The list_authorized_properties function enables buyers to confirm publisher authorization before purchasing inventory, addressing concerns about programmatic supply chain transparency.

Jason Widup, senior vice president of marketing at Pixis, described the protocol as "exactly the kind of infrastructure-level shift the advertising industry needs." Pixis, an AI-powered advertising solution provider, built its operating system on Model Context Protocol as a shared intelligence layer. According to a statement provided by Greenough Agency on October 15, 2025, Widup sees the protocol as "a way to reduce friction across fragmented tech stacks, unify signals between partners, and enable more automated, intelligent decision-making."

The timing aligns with broader industry developments in agentic AI adoption. McKinsey data indicates $1.1 billion in equity investment flowed into agentic AI in 2024, with job postings related to this technology increasing 985 percent from 2023 to 2024. Multiple advertising technology platforms have announced AI agent capabilities throughout 2025.

LiveRamp introduced agentic orchestration on October 1, 2025, enabling autonomous AI agents to access its identity resolution and audience activation platform. Adobe announced Experience Platform Agent Orchestrator on September 10, 2025, for managing agents across Adobe and third-party ecosystems. Amazon launched agentic AI capabilities on September 17, 2025, for autonomous marketplace management.

The protocol addresses challenges identified by industry analysts examining programmatic infrastructure. Traditional demand-side platforms require months of custom integration work, maintain fragmented reporting systems, and impose vendor lock-in through proprietary interfaces. The unified approach reduces integration timelines from months to days while providing consistent operations across platforms.

Model Context Protocol adoption has accelerated across advertising platforms throughout 2025. Google released an open-source MCP server for its Ads API on October 7, 2025, enabling AI applications to query advertising campaigns through natural language. Microsoft launched its Clarity MCP server on June 4, 2025, for web analytics queries. Google Analytics released its MCP server on July 22, 2025, while AppsFlyer introduced its MCP tool on July 17, 2025, for mobile marketing measurement.

Technical implementation requires publishers to expose their inventory through protocol-compliant servers. The reference implementation provides starting code for MCP server development. Validation test suites verify compliance with protocol specifications. Documentation covers data models, task definitions, error handling patterns, and webhook configurations.

AdCP supports signals activation workflows where audience data platforms expose targeting capabilities through standardized interfaces. Buyers discover available signals using natural language queries, evaluate pricing transparency, and activate selections on decisioning platforms. This enables direct integration between data providers and demand-side platforms without intermediary systems.

Industry adoption patterns remain uncertain. The protocol faces established competition from platform-specific APIs that major advertising systems have developed over years of investment. Network effects favor dominant platforms with extensive existing integrations. The success depends on whether enough publishers implement the protocol to create valuable inventory access and whether enough buyers adopt tools using the standardized interface.

Response times for complex operations reflect real-world advertising approval processes. Some publishers require legal review for new campaigns. Others enforce brand safety checks on creative assets. The protocol accommodates these workflows through its asynchronous design rather than attempting to eliminate necessary approval steps.

Ad Context Protocol does not address all advertising automation challenges. Fraud detection, viewability measurement, and attribution modeling remain platform-specific capabilities. Payment processing, invoice reconciliation, and financial reporting operate outside protocol scope. These functions require additional systems beyond the basic campaign management capabilities the protocol provides.

Implementation costs vary by organization size and technical capabilities. Small publishers may find MCP server development challenging without dedicated engineering resources. Large enterprises face integration complexity across existing technology stacks. The open-source approach reduces licensing costs but increases implementation burden compared to turnkey solutions.

The protocol's long-term viability depends on sustained community engagement and continuous maintenance. Open-source projects require ongoing contributor participation to address bugs, implement feature requests, and maintain compatibility with evolving underlying technologies. The founding members' commitment levels will determine whether the protocol achieves lasting industry adoption.

Timeline

• January 2025 – Perplexity AI founder outlines vision where AI agents become target audience for digital advertising
• June 4, 2025 – Microsoft launches Clarity MCP server enabling natural language analytics queries through AI assistants
• July 7, 2025 – Google explores MCP server for Ads API integration seeking developer feedback on implementation
• July 9, 2025 – StackAdapt launches AI assistant Ivy for programmatic advertising campaign management
• July 17, 2025 – AppsFlyer introduces MCP tool for mobile marketing measurement automation
• July 20, 2025 – Security vulnerabilities identified in MCP implementations affecting marketing platforms
• July 21, 2025 – Industry expert warns agentic AI threatens traditional DSP business models through automated campaign management
• July 22, 2025 – Google Analytics releases MCP server for AI-powered data conversations
• July 27, 2025 – McKinsey analysis reveals agentic AI leads frontier technologies reshaping marketing with $1.1 billion investment in 2024
• September 3, 2025 – Gracenote launches MCP server to verify AI responses for television platforms
• September 10, 2025 – Adobe launches Experience Platform Agent Orchestrator for managing AI agents across marketing workflows
• September 12, 2025 – Adverity debuts AI-powered intelligence layer built on Model Context Protocol technology
• September 17, 2025 – Amazon launches agentic AI capabilities for autonomous marketplace and advertising management
• October 1, 2025 – LiveRamp introduces agentic orchestration enabling AI agents to access identity resolution platform
• October 7, 2025 – Google releases open-source MCP server for Ads API enabling AI tools to query campaigns
• October 15, 2025 – Ad Context Protocol launches version 2.0.0 with 23 participating organizations

Summary

Who: Ad Context Protocol project led by Brian O'Kelley with six founding members—PubMatic, Scope3, Swivel, Triton, Optable, and Yahoo—plus 23 launch member organizations including AccuWeather, Adgent, Bidcliq, Butler/Till, Classify, HYPD, Kargo, Kiln, LG Ad Solutions, Locala, Magnite, Media.net, MiQ, Nativo, Newton Research, OpenAds, Raptive, Samba TV, Scribd, The Product Counsel, and The Weather Company.

What: Open-source advertising automation protocol providing unified interface for AI agents to manage campaigns across multiple platforms through standardized tasks covering product discovery, campaign creation, creative management, and performance optimization, built on Anthropic's Model Context Protocol with MIT licensing.

When: Launched October 15, 2025, with version 2.0.0 introducing production-ready capabilities including 9 core tasks, asynchronous operations, human-in-the-loop workflows, and support for both Model Context Protocol and Agent-to-Agent Protocol access methods.

Where: Global implementation through GitHub repository containing documentation, reference implementations, and validation test suites, with participating organizations spanning advertising technology platforms, publishers, data providers, and media buying systems across connected television, digital out-of-home, audio, display, and video channels.

Why: Addresses advertising ecosystem fragmentation where each platform maintains proprietary APIs requiring months of custom integration work, creating inefficiencies for media buyers and agencies navigating multiple systems, while enabling emerging agentic AI capabilities to automate campaign management across platforms through natural language interfaces and standardized workflows.