Document Intelligent Action: Document Intelligence action is required to extract the content of the documents and save them in a target location which is further used by copilot.
Pre-requisites:
Users generally need an Azure subscription to set up and use Azure services. User needs permission to create Azure resources and to deploy models for Azure Open AI.
Please refer below article to create Azure Services:
Instructions to create Azure Services - Crow Canyon Software Support
Copilot can work with different data sources; prerequisites vary based on the type of data source that is used to generate the answers.
AI Connections in NITRO Studio
Connections are a way to authenticate and consume both Microsoft and other resources within your Azure AI Studio projects.
Below are the supported connection types in NITRO Copilot.
Connection Type |
Description |
Azure Open AI |
Azure OpenAI Service provides REST API access to OpenAI's powerful language models including GPT-4o, GPT-4 Turbo with Vision, GPT-4, GPT-3.5-Turbo, and Embeddings model series. |
Open AI |
OpenAI is a cloud-based platform that allows developers to access a variety of models hosted by OpenAI, configure and submit prompts to said models and use their generative outputs in automations. |
Document Intelligence AI |
Document Intelligence is an AI service that applies advanced machine learning to extract text, key-value pairs, tables, and structures from documents automatically and accurately. |
Azure AI Search |
Azure AI Search is an Azure resource that supports information retrieval over your vector and textual data stored in search indexes. |
To configure the connections in NITRO Studio, go to the site à Site Settings à Crow Canyon NITRO Site Settings à Expand “AI Connection Settings” and configure the connections as shown below:
Document Intelligence AI Connection Settings
For document intelligence AI connection, users need Azure AI services multi-service account. API key and Endpoint value can be accessed from the corresponding resource in Azure portal.
Configuration of Document Intelligence Action
Library used: ‘HR Policies’ library in Helpdesk application
‘HR Policies’ library stores all the documents. Document Intelligence action is required to extract the content of the documents and save them in a target location. Target location can be ‘Current Item’, ‘Variable’, ‘Azure AI Search’ or a’ SharePoint List’.
-The following Workflow is configured on the ‘HR Policies’ library in the NITRO Helpdesk application item added event and item updated event.
Action settings:
General Settings:
Enter an action name, action description, and delay execution for the action
Content Settings:
Operations:
Add/Update Content:
▪In case of item added event
•This option adds the content from document library in target location for the copilot
▪In case of item updated event
•This option updates the content from document library in target location for the copilot
▪In case of item deleted event
•This option is disabled for item-deleted event
Remove Content: This option removes the content so that content is not used in copilot.
▪In case of item deleted event
•Select the Remove Content operation if the content is deleted
▪In case of item updated event
•content is still present in the library but is made unavailable for the copilot.
▪In case of item added event
•This option is disabled for item-added events.
Content Source: Select the source of the content, either a document or any multi-line text column. If Workflow is configured on a custom list, it will ask for a file name for the document option.
Document Intelligence Settings:
AI Connection: Select the Document Intelligence connection configured in NITRO Site Settings. If no connection is configured yet, click the plus ‘+’ icon to configure one.
Split Document: select ‘yes’ if you want to split the document. The ‘Chunk Size’ and ‘Chunk Overlap’ columns will be visible if the split document is selected.
Chunk Size: The entered number will show the maximum number of characters in each chunk of the document after splitting.
Chunk Overlap: It determines how much chunks overlap with each other. It is used to give continuity and preserve the meaning of text in the chunks.
Target Settings:
Select Target: here, you can select the target where you want to store the data extracted from documents.
Current Item: Use this option to extract the content from the document and put it in some multi-line text column of the current item for later use. This option will be disabled in the case of the Split Document option because It is not possible to store multiple document data in single target column. In current item, we can store data from one document in target column. Creating multiple columns to store multiple documents data is not a valid option as the target ‘Text column’ doesn’t support the selection of multiple columns.
Set in Variable: This option can be used in scenarios where the user wants to extract the data from the document and store it in a variable. The variable can then be used for example to send data by email by send mail action template. This option will be disabled in the case of the Split Document option.
Azure AI Search: Use this option to store the data extracted from documents in Azure AI Search as an index. When a user makes a query, NITRO copilot will use the index to give more relevant results.
SharePoint List: Use this option to store the data extracted from a document in a multi-line text column of a list.
Embeddings AI Connection: Select the Embeddings AI connection configured in NITRO Site Settings. If no connection is configured yet, click the plus ‘+’ icon to configure one.
Azure AI Search Connection: Select the Azure AI Search connection configured in NITRO Site Settings. If no connection is configured yet, click the plus ‘+’ icon to configure one.
Azure AI Search Index: Select the search index from the dropdown. If no search index is shown in the dropdown, then you can configure it in NITRO Copilot AI Settings Section.
-Configure workflow on the ‘HR Policies’ library in the NITRO Helpdesk application item deleted event. This workflow removes the content from the target location
- when the content is removed in the ‘HR Policies’ library.
For more information please refer Document Intelligence documentation - Quickstarts, Tutorials, API Reference - Azure AI services | Microsoft Learn