--- title: Generative AI API Endpoint Details slug: tutorials/generative-ai-api-endpoint-details docTags: createdAt: 2024-02-06T12:41:36.089Z --- Each of the Generative AI endpoints utilize the unique **`interactionIdentifier`** returned upon successful declaration of the [audio](docId\:PqLj1vqO7gTf0iqj1LBVN), [chat](docId\:iD41ifX7uC0RZDlb7RbtO), or [transcription](docId:8OMp8XUfyRY9gaB1YIvul) interaction. **NOTE:** [Audio](docId\:PqLj1vqO7gTf0iqj1LBVN) interactions must have completed processing before the Generative AI Endpoints can provide success responses. Attempting to make a request prior to the completed processing of an [audio](docId\:PqLj1vqO7gTf0iqj1LBVN) interaction will result in an error response. [Chats](docId\:iD41ifX7uC0RZDlb7RbtO) and [transcripts](docId:8OMp8XUfyRY9gaB1YIvul) allow for immediate access to Generative AI endpoints after declaration. The **Generative AI REST API** calls are all similar in format. For details on that format, please see the [Making API Requests for GenAI Results](https://docs.elevateai.com/tutorials/making-api-requests-for-generative-ai-results) page. Unsuccessful response codes are documented on that page, as well. Access a guide to using these features, below. # // CX Generative AI API Features ## AutoSummary - **Description: **Retrieves a summary of the interaction between the Agent and Customer - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/cx/summary** :::CodeblockTabs Successful Response ```json { "summary": string } ``` ::: ## Summary Details - **Description: **Summarizes additional details about the interaction, including the reason(s) for the interaction, actions taken, entities mentioned, and whether or not a resolution was reached - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/cx/summary/details** :::CodeblockTabs Successful Response ```json { "reasons": string[], "actionsTakenByAgent": string[], "entities": [{ "type": string, "value": string }], "isResolved": boolean } ``` ::: ## Agent Action Items - **Description: **Provides insight into the specific actions taken by an agent over the course of a customer interaction - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/cx/agent/action-items** :::CodeblockTabs Successful Response ```json [{ "action": string, "details": string }] ``` ::: ## Agent Coaching Assistant - **Description: **Provides insight into agent behaviors throughout an interaction, allowing users to identify opportunities for agent coaching and/or agent improvement after a customer interaction - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/cx/agent/coaching-assistant** :::CodeblockTabs Successful Response ```json { "coaching": string } ``` ::: ## Interaction Topics - **Description: **Provides a list of the topics that were discussed during the interaction - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/cx/topics** :::CodeblockTabs Successful Response ```json [{ "topic": string, "description": string }] ``` ::: ## Conversation Composition - **Description: **Provides a list of the questions and/or concerns that were raised by the customer during an interaction, the steps used to address and/or resolve each issue, and the status of each issue, post-interaction - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/cx/conversation-composition** :::CodeblockTabs Successful Response ```json [{ "concernSummary": string, "concern": string, "isResolved": boolean, "actionsTakenByAgent": string[] }] ``` ::: ## Ask ElevateAI Q\&A - **Description: **Allows the user to ask questions around a specific interaction – up to 5 questions per interaction – and receive specific answers. - The data being queried is your data! - A single question – up to 2000 characters – is asked in the request body of the **POST **method. The answer to the question is returned in the response. - Up to a maximum of 5 successful questions or 5 error responses are allowed per interaction. - This endpoint also supports a **GET** method which returns all of the prior question & answer pairs for this interaction (including failed questions and responses) - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/cx/ask** :::CodeblockTabs POST Request Body ```json { "question": "" } ``` POST Successful Response ```json [{ "question": string, "answer": string }] ``` GET Successful Response ```json [{ "question": string, "answer": string }, { "question": string, "answer": string }, { "question": string, "answer": string }] ``` ::: # // Generic Generative AI API Features (Non-CX) ## Summary - **Description: **Retrieves a summary of the interaction - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/summary** :::CodeblockTabs Successful Response ```json { "summary": string } ``` ::: ## Summary Details - **Description: **Summarizes the entities mentioned within the interaction - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/summary/details** :::CodeblockTabs Successful Response ```json { "entities": [{ "type": string, "value": string }] } ``` ::: ## Action Items - **Description: **Describes the action items which were raised during the interaction - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/action-items** :::CodeblockTabs Successful Response ```json [{ "action": string, "details": string }] ``` ::: ## Topics - **Description: **Returns a list of topics that were discussed during the interaction - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/topics** :::CodeblockTabs Successful Response ```json [{ "topic": string, "description": string }] ``` ::: ## Ask ElevateAI Q\&A - **Description: **Allows the user to ask questions around a specific interaction – up to 5 questions per interaction – and receive specific answers. - The data being queried is your data! - A single question – up to 2000 characters – is asked in the request body of the **`POST`** method. The answer to the question is returned in the response. - Up to a maximum of 5 successful questions or 5 error responses are allowed per interaction. - This endpoint also supports a **`GET`** method which returns all of the prior question & answer pairs for this interaction (including failed questions and responses). - **Endpoint: **/v1/interactions/\{interactionIdentifier\:guid}/**gen-ai/ask** :::CodeblockTabs POST Request Body ```json { "question": "" } ``` POST Successful Response ```json [{ "question": string, "answer": string }] ``` GET Successful Response ```json [{ "question": string, "answer": string }, { "question": string, "answer": string }, { "question": string, "answer": string }] ``` ::: **Need more help?** Contact the [ElevateAI Support team](mailto\:support@elevateai.com).