Test Center

Create test suites and run simulated calls.

The Test Center lets you validate agent behavior before publishing.

Concepts

Concept	Description
Test Suite	A set of test cases associated with an agent.
Test Case	An individual scenario with a prompt and success criteria.
Test Run	Execution of a suite with results per case.
Publish Readiness	Indicator whether the agent is ready for publishing.

Create a suite

Go to Test Center.
Click New suite.
Select the agent.
Choose language and whether to use the draft by default.

Create cases

Each case has:

Name: identifier.
Scenario prompt: a description of the simulated customer's role.
Success criteria: a list of criteria the call must meet.
Success mode: all (all must pass) or any (at least one).
Assert knowledge: checks about sources used.

Click Gerar casos (Generate cases) and pick scenarios (ready-made presets are available). The platform creates up to 25 cases per request with suggested criteria — each case accepts up to 20 criteria and simulated conversations run up to 30 turns.

Run

Click Run. The suite will run each case as a simulated call:

generates a transcript;
evaluates criteria;
returns passed/failed per case.

Run states

Run status (the suite execution):

State	Meaning
`queued`	Waiting for an execution slot.
`running`	Execution in progress.
`completed`	Execution finished; check per-case results.
`failed`	The execution failed technically.
`cancelled`	Execution cancelled.

Per-case status (each scenario inside the run):

State	Meaning
`queued`	Waiting to run.
`running`	Simulated call in progress.
`passed`	All criteria passed.
`failed`	At least one criterion failed.
`error`	Technical error while running the case.

Review

Open the run to see:

pass/fail summary;
full transcript;
results of each criterion;
flow diagnostics;
publish readiness.

When the agent passes tests, configure telephony and channels in Telephony.

Citations

Understand where the agent's response came from.

Evaluations

Success criteria and evaluation modes.