## AI Agents Learn to Work Together

This weekly update explores the exciting movement toward agent collaboration—where artificial intelligence systems work together seamlessly to accomplish complex goals. Throughout this week, major technology companies announced new partnerships and integrations that show how AI agents are becoming smarter at teamwork and coordination.

## Chemistry and Medicine Get Smarter AI Helpers

Evogene Ltd., a computational chemistry company, and Google Cloud announced an expanded partnership that marks a significant milestone in agent collaboration. The new phase focuses on developing and integrating advanced AI Agents into Evogene's ChemPass AI platform using Google Cloud's Vertex AI technology. These AI agents represent a new generation of artificial intelligence that can independently plan, reason, and execute complex, multi-step scientific workflows. Instead of scientists manually guiding each step, the agents can now automate key discovery processes and enable large-scale parallel molecular exploration. This means researchers can identify and optimize small molecules for drug and agricultural development much faster than before. The ability of agents to work autonomously while coordinating with other systems demonstrates how collaboration happens not just between people and AI, but between different AI systems as well.

## Building Better Infrastructure for Agent Teamwork

Nebius, a cloud computing company, is acquiring Tavily, a real-time search company, to create a more complete agentic platform. This collaboration means AI agents will have built-in abilities to navigate the web, verify facts, and execute complex tasks without customers having to connect multiple separate vendors together. By embedding Tavily's search technology into its cloud platform alongside top-tier NVIDIA GPUs, Nebius creates an ecosystem where vertical AI companies and enterprises can build, tune, and run agents with a native search layer. This is a perfect example of how agent collaboration works—instead of agents working in isolation, they now have access to shared infrastructure that helps them all perform better.

Another important development comes from Anthropic, which is extending its Model Context Protocol (MCP) with a new app framework. This framework allows developers to build full-stack agentic applications where AI agents can drive interactive, multi-step experiences. Rather than just being simple chatbots, these agents can now discover tools, manage state, and execute complex workflows. This means agents can collaborate not just with other agents, but with entire applications, creating richer and more maintainable AI systems.

## Real-World Enterprise Collaborations Taking Off

In the United States, DXC Technology has completed one of the largest enterprise deployments of Amazon Q, rolling it out to 115,000 employees across 70 countries. The company has created internal AI Advisor Agents used by more than 40,000 engineers. This shows how agents can collaborate within large organizations to help people work better. BMC has also signed a five-year agreement with AWS to deepen integration between BMC's intelligent automation and AWS services. This collaboration focuses on using BMC's AI advisor called Jett to modernize and automate data pipelines and workflows.

Experian integrated its Aperture Data Studio with Snowflake's AI Data Cloud, allowing customers to design data quality workflows without moving their data around. This type of collaboration between different companies' technologies helps agents work more effectively with enterprise data. Perforce Intelligence is extending MCP server capabilities across multiple development tools so AI agents and copilots can work inside existing DevOps workflows. Rather than replacing established toolchains, agents now integrate with them.

## AI Agents Transforming Sales and Commerce

Salesforce released its 2026 State of Sales report showing that AI agents have become the top growth tactic for sales teams. With 54% of organizations leveraging AI agents for deeper automation, agents are now coordinating research and content creation tasks across the sales cycle. Top-performing teams are 1.7 times more likely to use AI agents than underperforming teams. This demonstrates agent collaboration at work—agents helping human sales teams accomplish more together.

On the commerce side, Google released its Universal Commerce Protocol (UCP), developed in collaboration with over 20 global partners including payment networks and major retailers. This protocol enables AI agents to coordinate and execute transactions seamlessly. Similarly, Coinbase launched Agentic Wallets specifically designed for AI agents to transact securely. These developments show how agent collaboration extends beyond information processing to actual financial transactions.

## Why This Matters

These announcements reveal a fundamental shift: AI agents are no longer working in isolation. They are learning to collaborate with other agents, with human teams, with different software systems, and with infrastructure platforms. Companies like Google Cloud, Anthropic, and others are building the frameworks and protocols that make this teamwork possible. As these systems mature, the potential for AI agents to solve complex, multi-step problems grows exponentially.