v1

latestOpenAPI 3.0.02026-07-24202886.6 KB
Speech

Queues recordings for processing

post/speech/v0/bulk-ingest

Request body

call_identifierstring required

Unique record identifier, typically the Universal Call ID (UCID)

speech_file_namestring required

The name of the recording file as uploaded to Medallia Managed File Transfer

unit_identifierstring required

ID of the agent that handled the call (typically the last agent the customer is transferred to if there are multiple agents). This must match the ID that is included in the organizational hierarchy for the agent. If no value is provided, either a default unit or "Unit Pending" will be used for the record.

call_date_and_timestring date-time required

Interaction datetime in ISO-8601 format yyyy-MM-DD'T'HH:mm:ssZZ (e.g. 2021-03-22T11:30:00-08:00)

call_recording_urlstring uri

URL of the call interaction recording, referencing an external system for linkage purposes (not used for call transcription processing)

vertical_model'Call Center' | 'Financial Services' | 'Healthcare' | 'Voicemail' | 'Survey' | 'Large Vocab' | 'General'

Specific speech vertical model to use

localestring

Primary language spoken during the call, specified in a format compliant with IETF BCP 47. This value is applied to both the agent and customer channels. (See agent_locale for further details.) The engine selected may or may not leverage this setting during its processing; please check with your Medallia specialist as per the engine configuration.

agent_localestring

The language of the agent spoken during the call, specified in a format compliant with IETF BCP 47. If this is omitted, the locale language is used for the agent. If this is set, the agent_locale is used instead. This is only used when engine is set to "Engine 1".

apply_diarization'Yes' | 'No'

Applies diarization (channel assignment based on vocal audio properties) to single-channel recordings

agent_channel'0' | '1'

The audio channel (when recorded in stereo) associated with the agent

substitutionsobject

A set of substitutions to apply to the transcribed text, in key/value format, where the key will be replaced by the value in transcription.

oovobject

A set of custom vocabulary terms with optional soundouts. Accepts JSON dictionary and list formats, or a comma- delineated string of text. Custom vocabulary improves transcription for accuracy words that are not in the language model's standard vocabulary.

apply_redaction'Yes' | 'No'

Applies redaction to the audio and transcribed text

first_namestring

The first name of the customer.

last_namestring

The last name of the customer.

emailstring email

The email address of the customer.

phone_numberstring

The phone number of the customer, typically from ANI (Automatic Number Identification).

connection_idstring

The unique identifier of a specific connection profile. This value does not typically need to be set. Please consult with your Medallia Speech specialist to determine if a value is required. This field has been deprecated since Medallia Experience Cloud version 22CR3.0, please use connector_id instead.

profile_uuidstring

This field has been deprecated since Medallia Experience Cloud version 21CR3.0, please use connection_id instead.

engine'Engine 1' | 'Engine 2' | 'Engine 3'

The Medallia Speech engine to use to perform the transcription processing. The value to use is determined by your Medallia Speech specialist during the initial audio analysis process while onboarding to Medallia Speech.

connector_idstring uuid

The unique identifier of a specific Speech Connector API as configured in Medallia Admin Suite. If set, the Speech data in the API request is routed through the Connector for processing, including for extra metadata available in speech_additional_info. This field was introduced with Medallia Experience Cloud 22CR3.0.

speech_additional_infoobject

Additional metadata to add to the Medallia Speech record. Mapping this data to the record is performed in setting up the Medallia Speech API Connector in Admin Suite. Use of this field requires setting a connector_id.

Example request

[
  {
    "oov": {
      "Matias": [
        "Marisa",
        "Mateeas"
      ],
      "other": [],
      "Juampa": [
        "Juanpa",
        "Juan Pablo"
      ]
    },
    "speech_additional_info": {
      "FirstName": "Jo",
      "LastName": "Smith",
      "AddressInfo": {
        "City": "San Francisco",
        "State": "CA",
        "Country": "USA"
      },
      "CallSegmentData": [
        0.3,
        5.8,
        8.9
      ],
      "VIP": true
    }
  }
]

Response

Processing status of the submitted job request

job_idstring uuid

The job identifier generated by Medallia

status'ACCEPTED' | 'PARTIALLY_ACCEPTED' | 'REJECTED'

The job processing status, either ACCEPTED or REJECTED (to indicate that all recordings were processed with the same result) or PARTIALLY_ACCEPTED (to indicate a mixed result)