Cosign launches agent-to-agent protocol for professional networking
Cosign is introducing an agent-native interface that allows AI agents to negotiate on behalf of users within its professional networking platform. The move aims to replace traditional cold outreach with automated, first-pass evaluations of mutual goals.
Cosign, a professional networking platform, is developing an agent-native interface to enable AI agents to negotiate on behalf of users. The platform offers an 'Ask Cosign' conversational agent, an iMessage agent, and a personal Model Context Protocol (MCP). This MCP allows users to connect external agents, such as Muse, GrokBot, or Instinct, to the Cosign graph. This development follows early agent-to-agent communication protocols introduced by Decagon with Muse and Instinct for customer success.
The primary goal is to enable a user's personal agent to connect with a company's agent to perform a first-pass evaluation of mutual goals and intent. This process potentially bypasses traditional cold outreach methods. Cosign aims to provide a searchable data layer for the startup ecosystem, noting that current large language models provide poor answers for finding specific startups or talent. The company also highlighted that LinkedIn lacks an MCP for vast graph search.
David Booth and Dhruv Gupta argued that traditional hiring and networking methods are outdated. They contended that agent-to-agent communication will allow for more efficient matching of talent and opportunities based on trusted reputation signals. Booth illustrated the concept by asking what it would mean if a job seeker's agent could connect with a company's agent to determine whether it is worth elevating the candidate to a hiring manager.