Merge

Merge's Universal Context Layer is an enterprise feature inside Merge for Workforce that gives every AI tool or agent employees use a shared, permissioned view of company knowledge. It syncs and vectorizes data from company systems ahead of time so an agent queries a pre-indexed knowledge base instead of firing a live API call for every question, with connectors for file storage, ticketing, and HRIS at launch . The result is that a marketing hire opening Claude, or an analyst opening a different assistant, arrives with the same institutional context a colleague already taught into a separate tool, without re-uploading files or re-explaining internal systems. The launch matters because token spend has become a real budget line and most of it is wasted rediscovering things the company already knows. Merge frames runaway token cost as a symptom of a broader problem, where employees spend as much of their day setting up AI tools with the right context as they do actually using them to get work done . Because the context layer knows which tools and skills fit a task, an agent reaches for the right ones instead of probing tool after tool, and each request runs under the employee's own permissions so an agent sees only what that person would see . The layer follows a team to whatever AI tool or agent they adopt next, running with the same data access as the employee using it . Merge was founded in 2020 by Shensi Ding and Gil Feig , and is best known for its unified API for customer-facing integrations. This release extends that integration DNA inward, positioning Merge alongside its Unified, Agent Handler, and Gateway products as infrastructure for production AI covering context, tools, and models . For operators evaluating how to govern internal AI usage across a growing stack of assistants, it is a concrete alternative to per-tool prompt libraries and one-off retrieval setups.