Get started with Medallia Speech

Bring voice data into Medallia Experience Cloud for analysis and action.

Medallia Text Analytics is a key part of the Medallia Speech offering, since Text Analytics is used for revealing rich business insights from Speech transcription data. For this reason, Speech requires Text Analytics to be implemented on the company instance. If you need to perform that procedure, see Implement Text Analytics.

Complete the following steps to implement Speech:

Tip: Test the file-transfer process manually during the initial implementation to ensure the data is transferred correctly. After launching the program, you can fully automate the process.
  1. Contact Medallia Support to request a Medallia Media File Transfer (MMFT) storage bucket and credentials for your company.

    Your company will use the bucket for storing audio files to be transcribed by Speech. After confirmation that your bucket has been provisioned, you can find details for your bucket on the Medallia Media File Transfer screen in Setup (Integrations > Medallia Media File Transfer).

  2. On the OAuth screen in Setup, (Integrations > OAuth > Clients tab), configure a new OAuth client to connect with the Speech API.

    Configure the following properties with the specified values in addition to any other configuration for the client:

    • Confidential — Select the checkbox.
    • Grants — Include "Client Credentials Grant (for backend services)".
    • Client subject — Select "mec_provisioning".
  3. After saving the OAuth client, select it in the list of clients and click Generate new client secret. Share the Client ID and Client Secret with your company stakeholder.
  4. (Option 1) To pull call data files into Experience Cloud, configure one or more connectors to automate scheduled transfers of call data to your MMFT bucket.
  5. (Option 2) To use the Speech API to transfer audio files manually, complete the following steps:
    1. To include custom call metadata with the audio files, configure an Importer that appends that metadata to the call records in Experience Cloud.

      For example, custom call metadata might include wait time, hold time, number of holds, and extension number.

    2. If you configured an Importer for custom metadata, configure an import feed that references that Importer.

      The import feed can be a Web Feed or a Feed Pull.

    3. Instruct your company stakeholder to use the Speech API to trigger call file processing.
  6. Optionally, make one or more of the following revisions:

    • Make phonetic substitutions until company stakeholders have agreed on acceptable transcription quality. For more information, see Advanced features.

    • If needed, work with your company stakeholder to disable redaction. For more information, see Speech redaction syntax and default rules.

    • Configure Text Analytics custom lexicons (user features) to capture company-specific terms. For more information, see User features.

    • Review existing Text Analytics topics. If needed, build new topics to support the data you expect to capture through Speech. For more information, see Topics.

  7. As needed, build new reports and modify existing reports. For more information, see Reporting and data fields
  8. If needed, create new roles for the company. No specific permission is required to view and hear voice data in reports.

    To download audio files directly from reports with Speech data, roles need the "Download Speech Audio Files" reporting permission.

  9. If needed, create users for the people at your company, and assign those accounts to existing roles.
With these steps complete, you are ready to start processing call files with Medallia Speech!