1.
Engineer speaks the query into the mic — captured as raw audio
2.
ASR transcribes speech → text passed to the Chloe agentic orchestrator
3.
Orchestrator recalls prior context on this cell / session from Memory Store
4.
Orchestrator plans the use case: needs KPI tool + Fault/Inventory tool
5.
MCP Client sends JSON-RPC tool-call request(s) to the CHLOE MCP Server
6.
MCP Server queries Performance KPI, Fault & extended operator datasets
7.
Datasets return structured results to the MCP Server
8.
MCP Server returns the tool result on the return path
9.
Result relayed back through the MCP Client to the orchestrator
10.
Orchestrator synthesizes the answer and writes new findings to Memory Store
11.
Response text is sent to the Text-to-Speech engine
12.
Spoken answer is delivered back to the user
