Enterprise procurement teams are adding AI interoperability protocol support as a formal evaluation criterion when selecting new vendors, according to an analysis published by Influencers Time on July 13. The shift treats Anthropic’s Model Context Protocol (MCP), Google’s Agent2Agent (A2A) framework, and OpenAI’s tool-calling standards as table-stakes requirements rather than optional technical features.
The Procurement Shift
The change reflects a practical problem: most enterprise AI stacks now include five or more AI-powered tools from different vendors, and almost none share context natively. Each vendor has its own API conventions, its own data representations, and its own way of describing capabilities. When these systems need to work together, engineering teams build custom middleware that breaks every time a vendor pushes an update.
That custom integration work is expensive. Influencers Time reports that organizations spend up to 30% of their total software budget on integration and customization rather than the software itself, citing prior Gartner estimates. For AI-specific tools, which are newer and less standardized, that percentage often runs higher.
Interoperability protocols attack this cost directly. When vendors support common standards, integration overhead shrinks, time-to-value on new tools shortens, and switching costs drop. Procurement teams have started asking a single qualifying question during vendor evaluation: does this tool support open protocols, or does it lock us into proprietary integration?
What the Protocols Do
The three major protocols address different layers of the interoperability stack.
Anthropic launched MCP in November 2024 as an open standard for connecting AI assistants to external data sources and tools. It provides a universal interface for AI systems to access content repositories, business tools, and development environments. Early enterprise adopters included Block and Apollo, and development platforms like Zed, Replit, and Sourcegraph integrated MCP to let AI agents retrieve relevant context for coding tasks.
Google announced A2A in April 2025 with backing from over 50 technology partners including Salesforce, SAP, ServiceNow, Atlassian, and PayPal. A2A focuses specifically on agent-to-agent communication: letting AI agents from different vendors exchange information, coordinate actions, and collaborate across enterprise applications without custom bridges.
OpenAI’s tool-calling standards, while less formalized as a protocol, define how AI models describe and invoke external tools, creating a de facto standard that many third-party platforms have adopted.
Where This Matters for Agent Builders
The procurement shift has a direct implication for anyone building or deploying AI agents. Vendor lock-in is no longer just a pricing concern. It is a structural constraint on how agents can be composed.
Agentic workflows depend on fast, reliable context-sharing between systems. A media buying agent that needs to wait on manual data exports before adjusting spend, or a customer service agent that cannot access CRM data without a custom connector, defeats the purpose of running autonomous systems. When the protocols work, an agent managing programmatic ad spend can pull real-time performance data from a separate analytics tool without a custom API integration, because both speak the same standard.
The gap between promise and deployment is still real. None of these protocols are fully mature, and custom work will still be needed for unique business logic. But procurement teams are making bets on the direction of travel. Protocol support is becoming a line item that determines whether a vendor makes the shortlist.