MediaPost. What happens when AI agents collaborate in advertising?
MediaPost’s Laurie Sullivan recently wrote a piece about AI agents, and what happens to Google Ads when they do. AI agents are no longer science fiction – they’re already shaping how business gets done. And when they start talking to each other? That’s when the real transformation begins.
But there’s a catch: human language may not be ready for it.
Why AI agents aren’t ready to ‘talk’ to each other
In her commentary, Sullivan referenced one of 555vCTO.com founder Vaclav Vincalek’s newsletters: a weekly column called Recurrent Patterns. In it, he wrote about the intricacies of human language, and why AI is not ready
“Human language is suitable to describe the beauty of the world, and it lets our mind wander in various directions, but it is [too] ambiguous,” says Vaclav Vincalek, founder of Hiswai.com and long-time fractional CTO at 555vCTO.com.
He went on to explain that that ambiguity becomes a problem when you're building multi-agent systems that need to interpret precise instructions – or enforce contractual obligations – without human input. If two AI agents are negotiating ad placements or executing real-time bids, the last thing you want is a misunderstood “terms and conditions.”
AI agents are now a system and not a feature
The MediaPost article does mention the benefits of AI agents. Google’s new Agent2Agent (A2A) protocol is a blueprint for that system: a way for AI agents to collaborate across apps, data sets, and organizations. It’s like giving APIs intelligence – enabling agents to initiate and manage complex workflows on their own.
This is the foundation for the next evolution of advertising. In the future, agents could autonomously:
Buy and optimize ad placements
Adjust bids based on real-time inventory
A/B test creative and shift budget in seconds
No dashboards. No campaign managers. Just instructions – and the results that follow.
SSI and Google’s next AI bet in the AI agents game
Google knows this is the future. The article mentions how the tech giant has backed Safe Superintelligence (SSI), a startup co-founded by former OpenAI scientist Ilya Sutskever, now valued at $32 billion. Google’s cloud arm is also providing SSI with access to its in-house TPUs – Google’s own chips designed for AI workloads.
Sullivan explains that the investment likely wouldn’t have happened if the DOJ had succeeded in forcing Google to divest from AI companies like Anthropic. For Google, owning a stake in the agent-based future is a strategic move – especially as rivals like Microsoft and OpenAI push forward.
From Google Ads to AI agent-based advertising
So what happens to Google Ads?
Right now, it’s a human-led, platform-driven system. In the agent-to-agent ecosystem, campaign decisions are made in real time by AI.
The article writes that instead of building dashboards for marketers, Google may end up building infrastructure for agents – the very same agents that companies like Salesforce, Accenture, and Deloitte are already helping define through A2A.
And they’re not alone. Anthropic’s own Model Context Protocol (MCP) is a competing approach that allows agents to connect directly to tools and data. Think of it like Zapier, but for autonomous AI.
Is your startup developing an AI agent?
AI agents that talk to each other will upend many other sectors, not just the advertising industry.
Is your startup developing an AI agent (or thinking about it?). Reach out to the experts at 555vCTO.com to create a product roadmap to give your AI agent the most effective chance at operational efficiency and growth for your enterprise.
Let’s talk AI agent (or any other) business strategy. Contact the tech experts at 555vCTO.com today.