Customer Agent: Termination Handling
Tracks occurrences of the agent freezing, hanging up abruptly, crashing, or cutting off conversations early.
Customer Agent Termination Handling checks whether a conversation ends cleanly or cuts off abruptly, before the user’s issue was addressed. Run it to catch dropped calls, crashes, and premature endings.
What it does
Customer Agent Termination Handling is an LLM-as-Judge eval. It reads the full conversation and scores how often termination issues occur.
Input
| Required Input | Type | Description |
|---|---|---|
conversation | string | The full conversation history between the customer and agent |
Output
| Field | Type | Description |
|---|---|---|
| Result | never / occasionally / frequently / always | Indicates how often termination issues occur |
| Reason | string | A plain-language explanation of the termination handling assessment |
Run it from code
Call evaluate() with the template name and the eval’s required inputs. It returns the score and the reason.
Note
Before running: install the SDK and set FI_API_KEY / FI_SECRET_KEY. The model argument in the snippets is the evaluator model Future AGI uses to run the eval; turing_flash is a fast default.
from fi.evals import evaluate
result = evaluate(
"customer_agent_termination_handling",
conversation="User: I need help with my subscription.\nAgent: Sure, I can help with that. What would you like to know?\nUser: How do I cancel?\nAgent: [call ended]",
model="turing_flash",
)
print(result.score)
print(result.reason)import { evaluate } from "@future-agi/ai-evaluation";
const result = await evaluate(
"customer_agent_termination_handling",
{
conversation: "User: I need help with my subscription.\nAgent: Sure, I can help with that. What would you like to know?\nUser: How do I cancel?\nAgent: [call ended]"
},
{ modelName: "turing_flash" }
);
console.log(result);When to use
Run Customer Agent Termination Handling on voice and chat agents where a dropped session means a lost or angry customer.
- Voice agents, to catch calls that end abruptly mid-request
- Chat sessions that close before the user’s question was resolved
- Post-incident review, to check whether a spike in complaints traces back to abrupt endings
What to do when Customer Agent Termination Handling fails
Investigate system logs for errors or timeouts around the flagged moments, and add graceful error handling to prevent abrupt disconnections. Implement conversation state persistence so sessions can recover from crashes instead of dropping entirely.
Test edge cases that may trigger unexpected termination before they show up with real users.
Questions & Discussion