Symptom or Need

You detect that the activity code used to classify a call does not match the code reflected in the Speech Analytics report.

Context / Scenarios

• Affected Services or Modules: Reports Module, Data Monitor and Speech Analytics Report.
• Typing Methods: Operations that assign activity codes using manual Coding options, Scripting automations, or API injections.
• Interaction flow: Environments where the call process interacts with your databases or external APIs are consumed .
• Scope: The error may occur globally in the calls of the operation or partially in a specific percentage of calls triggered by the speech .

Answer / Solution

To manage this incident, we suggest you perform the following checks in your environment and gather the required information before submitting a support ticket to the service desk:

Initial Validations

1. Check how the calls are being typed in the operation (Coding, Scripting, or API).
2. Determine the overall impact percentage and the percentage of those calls that are activated by Speech .

Information to send to the support team

Generate the ticket for the service desk, attaching:

1. Name of the affected operation .
2. Percentage of overall impact of the failure.
3. Percentage of calls that are activated by Speech.
4. Names and numbers of the specific reports that do not match.
5. IDs, date and time of the affected calls , so that the technical team can validate if the information matches between the monitor and Speech Analytics .

Possible Causes

• Overwriting by integrations: When performing processes that involve the consumption of external APIs or databases during the call flow, the system may be overwriting the original encoding sent to Speech Analytics.

Considerations

• With the information provided, the support team will be able to run a controlled test by coding via API , in order to validate whether the system overwrites the coding under that method.

Practical Example

• Illustrative scenario: An agent ends an interaction and uses the manual option to "Code" the call as Soporte Técnico . However, during this closure, the customer flow consumes an API to update their CRM. This API returns an automated action that overwrites the code to Gestión Cliente . As a result, the data monitor displays Soporte Técnico , but Speech Analytics reads the last data injected by the API ( Gestión Cliente ), generating the discrepancy.