Stacks/Agent platform / infrastructure
Workday's internal Agent Platform
Workday's Agent Platform team, starting with four people, built the server, eval infrastructure, and an SDK that wraps LangChain primitives around Workday's LLM gateway and RAG service, with LangGraph for orchestration and LangSmith alongside. Prototype time fell from 4 to 5 months to under a day, and the team has grown to roughly 30. This is a LangChain customer story.
- Use case
- An internal developer platform for building, deploying, and running AI agents for HR, finance, and IT on Workday's infrastructure, with auth, security, compliance, and audit built in
- Category
- Agent platform / infrastructure
- Team size
- Started with 4 people; now roughly 30
- Orchestrator / framework
- Product teams build agents as LangGraph graphs, from simple ReAct loops to deterministic workflows, deployed on LangChain's Agent Server (A2A- and MCP-compliant). An internal Python SDK wraps LangChain primitives and routes them through Workday services: an LLM gateway, a RAG retrieval service, and other tooling. The Agent System of Record governs which APIs agents may call.
- Memory
- not stated
- Evals
- A collaborative eval platform built by the Agent Platform team, which the Agent System of Record runs on.
- Observability
- LangSmith.
- Cost (stated in writeup)
- Not stated
- Cost per output (stated)
- Not stated in the writeup
- WhatWeRun estimate
WhatWeRun estimate
Not enough info to estimate
The post doesn't name models or give any workload numbers.
Reviewed Nov 13, 2026
- Observability tools
- Other tools
- Agent System of Record (API governance)LLM gatewayRAG retrieval serviceA2A and MCP
- Source
- https://www.langchain.com/blog/how-workday-built-an-enterprise-scale-ai-agent-platform-with-langgraph-and-langsmith
Sofia Sulikowski (LangChain customer story) · Published Nov 10, 2026 · Last verified Nov 13, 2026
Related
More like this
developer.nvidia.comNVIDIA Agent Substrate
A security-hardened, zero-trust Kubernetes runtime for running sandboxed agents and tools at high density — coding assistants, RL post-training environments, and short-lived GPU sandboxes — now donated to the CNCF Sandbox
Datadog's Insights agent, tuned by its own traces
The agent behind Datadog Agent Observability Insights, which analyzes agent traces to find cost and reliability problems; Datadog uses Insights on that agent's own traces to cut tokens and fix failing tool calls
langchain.comNuman's Nu AI health coach
An AI health coach for patients on Numan's weight-loss and long-term-condition programmes, kept inside a UK regulatory line (coaching and education, no diagnosis or prescribing)