Skip to main content

Documentation Index

Fetch the complete documentation index at: https://koreai.mintlify.app/llms.txt

Use this file to discover all available pages before exploring further.

Back to Reports List The Latency Report highlights calls exceeding a defined latency threshold. For example, supervisors can generate a report of all calls where the P90 end-to-end latency is greater than 500 milliseconds (ms).
This is a beta feature. To generate Latency reports, enable capturing latency metrics by configuring these call control parameters:notifySttLatency = truevadEnable = truevadVendor = silero

​
Generating the Report

  1. On the Reports tab, select New Report.
  2. In the list of reports, select Latency Report.
  3. Enter the Report Name and Report Description.
  4. Select a Time.
  5. Select Data Filters.
    P90 ValuesThreshold Values (ms)
    P90 End-to-End LatencyFor example, 200 ms
    P90 ASR LatencyFor example, 200 ms
    P90 Bot Response TimeFor example, 200 ms
    P90 TTS LatencyFor example, 200 ms
    Each P90 value is evaluated using the following conditions: Greater than or equal to (≥), Less than or equal to (≤), Greater than (>), Less than (<), Equal to (=).
  6. Select Report Generation and choose Run On Demand or Schedule, where you can create up to six schedules (only one for the Hourly option).
  7. Select the Format (CSV).
  8. Enter a Send To address (single email address or mailing list).
  9. Select Create to generate the report.
  10. Select the Run option to generate the report. Run On Demand generates the report immediately, while Schedule configures automatic generation at the specified date and time without manual execution.
  11. Select the option to download the report.
The red dot beside the report indicates it was generated within the last 24 hours. You can download, save, and view the report in CSV format.
The report is available for download. Latency Report Output

​
Report Fields

FieldDescriptionExample
Session IDUnique identifier for each interaction session.a12b34c56d
Start TimeTimestamp when the session started.2025-09-24 10:56:34
AgentName of the agent handling the session.John Doe
Agent EmailEmail address of the agent.john.doe@example.com
Queue NameName of the queue the interaction routes through.Support_Queue
CustomerIdentifier or name of the customer involved in the session.Jane Doe
DurationTotal time the session lasted.00:12:45
P90 ASR Latency (ms)90th percentile latency of Automatic Speech Recognition, in milliseconds. Indicates how long speech-to-text processing takes.250
P90 Bot Latency (ms)90th percentile latency for the bot’s response processing, in milliseconds. Shows system responsiveness.120
P90 TTS Latency (ms)90th percentile latency of Text-to-Speech processing, in milliseconds. Represents the time to convert text into audio.180
P90 End To End Latency (ms)90th percentile overall latency from customer speech to agent/bot reply, in milliseconds. Combines ASR, bot, and TTS processing times.600