Skip to main content

Overview

This agent investigates anomalies in rocket engine hot fire tests. It synthesizes data across telemetry, hardware logs, inspection reports, and manufacturing records to identify root causes—work that would normally take engineers hours of cross-referencing disparate systems. Built with Agent Composer, Contextual AI’s new framework for production-ready AI agents.

Watch the Demo

The Scenario

Test 142: ME-1 Engine Static Fire A 15-second hot fire test of a methane/LOX engine just completed. The results:
  • Chamber pressure 4.2% below prediction
  • Thrust 3.8% low
  • No redline violations, no abort—just underperformance
The engine ran fine on Test 141 two weeks ago. What changed?

What the Agent Does

Ask the agent: “We just completed TEST-142 on ME-1-002. Chamber pressure came in about 4% below prediction. What’s going on?” The agent connects the dots across:
  1. Telemetry showing low fuel flow and high mixture ratio
  2. Hardware change logs showing an injector was replaced between tests
  3. Inspection reports explaining why the original injector was swapped
  4. Manufacturing records revealing the replacement injector’s orifice diameter is at the low end of tolerance
  5. Historical data showing this manufacturing lot has had similar issues
The agent surfaces the root cause: a fuel injector flow restriction caused by a replacement part from a different manufacturing lot.

Output

The agent generates a root cause assessment with:
  • Primary root cause identification
  • Supporting evidence from each data source
  • Citations to source documents
  • Recommended next steps

Try the Demo

Launch Demo

Investigate the anomaly yourself

Learn More