RingSense

RingCentral RingSense transcribes your voice calls and provides in-depth conversational insights, offering detailed analytics derived from post-call analysis of recorded sessions. Based on the richness of the conversation content, the insights may include the following:

• Transcript - Transcription objects with identified speakers and the timestamps of their utterances.
• Summary - Summary objects with AI-generated summary paragraphs and the timestamps of the conversations.
• HighLights - Highlight objects contain speaker utterances that capture key points, valuable insights, and notable moments for quick review and analysis.
• NextSteps - Next steps objects contain AI-generated action items for the call participants.
• AIScore - A numerical value that represents the confidence level of the RingSense AI's analysis and prediction.
• BulletedSummary - A list of AI-generated sentences summarizing the conversations in bullet points.

To learn more about RingSense and how to enable the service for a RingCentral RingEX account, please refer to this online article.

RingSense API

The RingSense API allows you to programmatically access RingSense data from your account.

The API functions at the account level and requires the "RingSense for Sales - Access Insights" user permission. This means that any user extension with this permission can access RingSense data for any user holding a RingSense license.

How to check if an extension has the permission to access RingSense data via API?

Check from the account admin portal

• Log in to your account admin portal using a super admin user.
• Select the 'Users' tab then click the 'Roles' option
• Select the role which is assigned to the user extension
• Scroll to the Artificial Intelligence section and check if the checkbox is checked

Check from the platform API

Authenticate the app using the login credentials of the user who is authorized to access RingSense data. Then, make a GET request to the API below:

/restapi/v1.0/account/~/extension/~/authz-profile/check?permissionId=ReadRingSenseInsights

If the user has the permission, the response will contain the info below:

{
  "uri" : "https://platform.ringcentral.com/restapi/v1.0/account/80964XXXX/extension/6228832YYYY/authz-profile/check?permissionId=ReadRingSenseInsights&targetExtensionId=6228832YYYY",
  "successful" : true,
  "details" : {
    "permission" : {
      "uri" : "https://platform.ringcentral.com/restapi/v1.0/dictionary/permission/ReadRingSenseInsights",
      "id" : "ReadRingSenseInsights",
      "assignable" : true,
      "readOnly" : false,
      "siteCompatible" : "Independent"
    },
    "effectiveRole" : {
      "uri" : "https://platform.ringcentral.com/restapi/v1.0/dictionary/user-role/1",
      "id" : "1"
    },
    "scopes" : [ "AllExtensions" ]
  }
}

There are two methods to access RingSense data: through push notifications (recommended method) or via a REST API call.

RingSense event notification

You can get notified when a new voice call recording analysis is completed via RingCentral push notification service. The RingSense data is included in the event payload together with the event metadata.

To receive RingSense event notification for voice call recordings, subscribe for this event.

Remember, the notification events are triggered only if a voice call was recorded by users who hold a RingSense license.

Read RingSense data

You can read the RingSense data of a recorded voice call if you know the sourceRecordId or the sourceSessionId. Both sourceRecordId and sourceSessionId values are included in the RingSense event payload.

Read RingSense data using a recording Id:

GET /ai/ringsense/v1/public/accounts/~/domains/pbx/records/[sourceRecordId]/insights

This endpoint retrieves the RingSense data for a specific call recording, identified by the sourceRecordId.

Read RingSense data using a telephony session Id:

GET /ai/ringsense/v1/public/accounts/~/domains/pbx/sessions/[sourceSessionId]/insights

This endpoint returns the RingSense data for all call recordings associated with the specified telephony session. Multiple call recordings can exist for a single session in scenarios such as call transfers between agents, where each leg of the call is recorded separately. The RingSense data for each individual recording will be included in the records array in the API response.

Example of API Response

{
    "title": "Call from John Smith to Tracy Debolah",
    "domain": "pbx",
    "sourceSessionId": "s-a0e7b361acc0az191bec7e275z7cae070000",
    "sourceRecordId": "2728084995017",
    "callDirection": "Outbound",
    "ownerExtensionId": "6228832XXXX",
    "recordingDurationMs": 125963,
    "recordingStartTime": "2024-09-04T20:44:16.390Z",
    "creationTime": "2024-09-04T21:15:43.597Z",
    "lastModifiedTime": "2024-09-04T21:15:43.591Z",
    "speakerInfo": [
      {
        "speakerId": "p-a0e7b361acc0az191bec7e275z7cae070000-1",
        "name": "John Smith",
        "accountId": "809646XXX",
        "extensionId": "6228832XXXX"
      },
      {
        "speakerId": "p-a0e7b361acc0az191bec7e275z7cae070000-2",
        "name": "Tracy Debolah",
        "accountId": null,
        "extensionId": null
      }
    ],
    "insights": {
      "Transcript": [
        {
          "end": 14.52,
          "text": "Okay, I record this meeting so we will have the RingSense meeting insights.",
          "start": 2.2,
          "speakerId": "p-a0e7b361acc0az191bec7e275z7cae070000-1"
        },
        ...
        {
          "end": 40.44,
          "text": "And I have some action items for myself.",
          "start": 36.6,
          "speakerId": "p-a0e7b361acc0az191bec7e275z7cae070000-1"
        },
        {
          "end": 46.12,
          "text": "As I have to complete the test.",
          "start": 42.52,
          "speakerId": "p-a0e7b361acc0az191bec7e275z7cae070000-1"
        },
        ...
        {
          "end": 125.88,
          "text": "Goodbye.",
          "start": 125.4,
          "speakerId": "p-a0e7b361acc0az191bec7e275z7cae070000-1"
        }
      ],
      "Summary": [
        {
          "start": 0,
          "end": 1,
          "value": "The call is about discussing the tasks and timeline for completing a project."
        }
      ],
      "NextSteps": [
        {
          "start": 42.52,
          "end": 53.32,
          "value": "Complete the test with own code for the API.",
          "speakerId": "p-a0e7b361acc0az191bec7e275z7cae070000-1"
        },
        {
          "start": 56.04,
          "end": 64.92,
          "value": "Write a developer guide for the API.",
          "speakerId": "p-a0e7b361acc0az191bec7e275z7cae070000-2"
        },
        ...
        {
          "start": 82.92,
          "end": 99.56,
          "value": "Publish the developer guide after review and editing.",
          "speakerId": "p-a0e7b361acc0az191bec7e275z7cae070000-1"
        }
      ],
      "AIScore": [
        {
          "value": "8"
        }
      ],
      "BulletedSummary": [
      {
        "value": "Agent confirmed they were recording the call for sentiment analysis purposes."
      },
      ...
    ],
    "CallNotes": [
      {
        "value": "Call notes.\nAgent expressed frustration over internal communication issues and ..."
      }
    ]
  }
}

SpeakerInfo

The speakerInfo is a list of objects, with each object contains the following key/value pairs:

Insights

The insights is an object that may contain the following objects:

When processing the insights objects—such as reconstructing conversations using the utterances in the Transcript objects, you can match the speaker ID from a transcript object with the corresponding speaker ID in the speakerInfo object. This allows you to display the speaker's name (if available) alongside the utterance text.