AGENTIC PROTOCOLS The tools — MCP connects an AI to the systems you already run: your CRM, your image library (DAM), and your ad platform. YOUR TOOLS CRM DAM — images Ad platform MCP — the Model Context Protocol: an open standard from Anthropic that lets any AI request data and act through your tools. MCP The agents — four specialized AIs, each calling tools via MCP, hand work down the chain from insight to optimization. AGENT CHAIN INSIGHTS AI MCP CREATIVE AI MCP CAMPAIGN AI MCP OPTIMIZER AI A2A A2A — Agent-to-Agent from Google Cloud: a shared language so agents from any builder coordinate tasks and feed results back. A2A The output — Dynamic Creative Optimization: a personalized ad assembled and shipped to each user at machine speed. DCO AD one per user MCP talks to tools, A2A talks to agents — DCO at machine speed.
Agentic Advertising

MCP. A2A. OMG. The New ABCs of Advertising AI

· 4 min read · Originally on LinkedIn
The gist

Two new standards reshape programmatic: Anthropic's Model Context Protocol lets a single AI pull from your CRM, asset library, and ad platforms, while Google's Agent-to-Agent protocol lets specialized agents coordinate as a team. Together they make dynamic creative optimization run at machine speed — insight, creative, delivery, and budget tuning handled by AIs that talk to tools and to each other.


Another Day, Another Acronym: Why MCP and A2A Are the Future of Programmatic Advertising (Yes, Seriously!)


Let’s face it - programmatic advertising has more three-letter acronyms than a government agency. DSP, SSP, DMP, CDP… and now, MCP and A2A? At first glance, they sound like new airport codes or the next boy band duo. But stay with me - these two may be the missing puzzle pieces that bring automation, personalization, and real-time optimization together in ways we’ve only dreamed of.

Let’s break it down, simply and clearly.

MCP: Your AI’s New Best Friend

First up, MCP. It stands for Model Context Protocol, and it’s like a magic key for your AI. Created as an open standard by Anthropic, MCP connects your AI to the tools you already use like your CRM, image library, or ad platforms. Think of it as a universal charger for your AI, whether it’s GPT, Claude, or something custom-built.

Here’s how it works in four easy steps:

  • Your AI says, “I need some info!” Maybe it wants a user’s location or a product image.
  • A helper called the MCP client sends that request to an MCP server using a simple format.
  • The server talks to your tools - like your CRM or Google Ads - and grabs the info.
  • The AI gets the info back and uses it to create an ad or make a smart move.

It’s fast, safe, and works with any system. Cool, right?


A Real-Life Example: Dynamic Creative Optimization (DCO)

Let’s make this real with Dynamic Creative Optimization, or DCO - a fancy term for creating ads that fit each person perfectly.

Picture this: your AI checks a user’s history in your CRM using MCP. It sees they love hiking gear. Then, it pulls a photo of your newest hiking boots from your image library - Digital Asset Management (DAM). In a flash, it writes a catchy line: “Love hiking? Try our new trail boots!” And boom — it sends the ad straight to your ad platform to show the user.

No more slow tagging or manual work. MCP makes it quick and personal, every time.


A2A: When Your AIs Team Up

Now that your AI can talk to tools, what if you have lots of AIs working together? That’s where A2A comes in. Short for Agent-to-Agent, A2A is a new idea from Google Cloud. It’s a way for different AI helpers to chat with each other, no matter who built them - Google, Salesforce, or you.

Here’s how A2A works:

  • Each AI has a “skill card” to show what it can do, like a mini-resume.
  • They follow a simple plan: suggest a task, do it, and finish it.
  • They talk securely using common tech like HTTP.

Why does this matter? Imagine one AI finding your best customers, another creating an ad, and a third tracking how the ad does. A2A lets them work as a team, sharing info and making decisions together.


MCP + A2A: The Dream Team in Action

MCP helps one AI talk to tools. A2A helps many AIs talk to each other. Together, they’re unstoppable. Let’s see them work on an ad campaign:

AgentProtocolWhat it does
Insights AIMCPGrabs user data from your CRM — like who loves hiking.
Creative AIMCPWrites an ad and picks a hiking boot image.
Campaign AIMCPSends the ad to your ad platform.
Optimizer AIA2AChecks the ad’s results, tweaks the budget, and tells the team what to do next.

It’s like having a whole ad team in your pocket - except they’re all AIs, and they never sleep!


Why You’ll Love This as a Marketer

Here’s what MCP and A2A mean for you:

  • Super Fast: Launch ads in seconds, not days.
  • Big Scale: Make thousands of ad versions without stress.
  • Super Personal: Create ads that feel made for each person.
  • Less Work: Let AIs handle the boring stuff while you stay in charge.

Your AI isn’t just making ads anymore - it’s running your whole ad strategy.


This Isn’t Just a Trend

MCP and A2A might sound like more tech jargon, but they’re a big deal. They’re not just about using AI - they’re about making AI work better for you. As DCO becomes a must-have, MCP and A2A make sure your AI isn’t just smart it’s connected and ready to act.

In today’s world, people’s attention is hard to get, and they expect a lot. With MCP and A2A, you can deliver ads that grab attention and feel just right. That’s not a nice-to-have it’s the new standard.

P.S. What’s Next?

What new adtech acronyms do you think we’ll see soon? Maybe AIA (Artificial Intelligence Assistant) or OMC (Omnichannel Magic)? Share your guesses below — I’d love to hear them! Just don’t make up too many new ones, okay?

THE NEW ABCs MCP — Model Context Protocol. An open standard that connects one AI to the tools you already use, like a universal charger: it reaches the CRM, the DAM image library, and the ad platform to grab info and act. MCP · MODEL CONTEXT PROTOCOL One AI talks to tools AI CRM DAM image library ad platform A2A — Agent-to-Agent. A way for different AIs to talk to each other: each shows a skill card, then they suggest, do, and finish a task together. The team: Insights, Creative, Campaign, and Optimizer. A2A · AGENT-TO-AGENT AIs talk to each other Insights AI Creative AI Campaign AI Optimizer AI DCO — Dynamic Creative Optimization. The payoff: ads that fit each person perfectly, assembled and shipped at machine speed. DCO at machine speed Dynamic Creative Optimization — an ad for each person MCP connects one AI to its tools; A2A connects the team.