back_office_ops · saas · workflow
Exa builds production multi-agent deep research agent with LangGraph
As Exa's architecture evolved from a simple answers endpoint to a more complex agentic deep research system, they needed a coordination framework — their original answers endpoint required no framework but could not support the new multi-agent complexity.
How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · User submits research query
A user submits a research query to the deep research agent, which autonomously explores the web to find structured information.
Tools used
LangGraphLangSmithExa API
Outcome
Exa built a production-ready multi-agent deep research system that processes hundreds of research queries daily, delivering structured results in 15 seconds to 3 minutes depending on complexity, with LangSmith observability informing production pricing models.
Results
Time savedhundreds of research queries daily
Grounding & classification
Source type: platform led case
18 fields verified against source quotes.
agentic workflowmulti agent workflowragknowledge basemetric backednamed customerproduction runtime claimedtools describedvendor confirmedworkflow describedsoftwarethroughput increaseplatform led caseback office opsagentic task execution