Conformity
—
Awaiting assessment
Signing in…
Consistency
Understand DASH behavior, one piece of evidence at a time.
Assessed sample
0
No assessed conversations yet
Confidence interval
—
95% Wilson interval · assessed sample
Awaiting assessment
0
Execution alone does not establish quality
Conformity over time
Happy pathAssessed results · selected target version and scenario contract
No assessments for this version yet.
What these numbers tell you
Run a battery, then ask your agent to assess the saved evidence.
The executor never assigns a quality verdict. A confidence interval describes this evaluated sample; it is not a guarantee across all users or scenarios.
Recent batteries
| BATTERY | EXECUTOR | EXECUTION | ASSESSMENT | STARTED | EVIDENCE |
|---|
Run tests
Watch the AI follow the scenario in the browser, with every step recorded.
Configure battery
Settings are fixed once the battery starts.
Battery queue
AgentCore Browser
The actual session where the AI operates the authenticated widget.
Offline
The browser will appear hereStart a battery to follow the AI's actions live.
Live DCV stream · view only · actions remain controlled by the scenario.
Battery progress
0 / 0 completedWaiting for a battery.
Observed conversation
Negative feedback
The exact message, with the context needed to understand what happened.
New · loaded items
0
In review · loaded items
0
Addressed · loaded items
0
Feedback has not been loaded.
Inbox
Conversation context
Select a message to view its context.
Channel parity
Chat scenarios match the Run tests selector. Is each scenario implemented in Webcaller and PSTN? Consistency tracks flow maturity and regressions; voice validation is a future decision.
ImplementedNot implemented— Not applicable (excluded)
| CHAT SCENARIO (REFERENCE) | WEBCALLER | PSTN |
|---|---|---|
| Loading parity manifest… | ||