Customizing Copilots with own data Using Azure AI search, Azure AI studio, and Copilot Studio.

Copilot Studio is a Microsoft platform that allows you to create and customize AI copilots for various applications and services. Here are some key features:

• Low-Code Development: You can build and customize copilots using a graphical interface, making it accessible even if you don’t have extensive technical skills¹.
• Integration with Microsoft 365: Copilot Studio allows you to enhance Microsoft 365 applications like Teams, SharePoint, and Outlook by adding custom AI capabilities².
• Automation and Workflows: You can design workflows and automate tasks using Power Automate and other tools, making your copilots more efficient and responsive¹.
• Data Connectivity: The platform supports connecting to various data sources, both within and outside of Microsoft 365, to provide rich, dynamic responses².
• Multi-Channel Deployment: You can deploy your copilots across different channels, including websites, mobile apps, and social media¹.

Custom copilots can be created by using generative AI features as well. If copilots are to be created based on different LLM models and different data sources, Azure AI Studio can be used.

Data can be stored in Azure Blob storage, and vectorization can be done using Azure AI search.

Creation of a Storage Account

The storage account can be used to store the required data sources on the blob storage.

Provision of Azure Open AI Studio

The GPT models can be deployed on Azure AI studio using this resource.

Provide the mentioned parameters and the resources.

Click on Create and Provision of Azure Open AI Studio.

Provision of Azure AI Search resource

Provision of Azure AI search service in the same region using the same resource group. The Azure AI search can index the data and be added as the data source in AI Studio.

Upload the grounding data to the container by clicking on the storage account and upload. Create a new container or select from the existing container.

Deployment of models on Azure AI studio

Create two deployments in Azure Open AI studio.

i) A GPT model that generates responses based on the grounding data.

ii) A text embedding model for vectorization using Azure AI search to provide grounding data.

Step i) In Azure Open AI studio, select deployments and select the Azure based model for deployments.

Select a model based on the application requirement.

Deploy the GPT-35-turbo-16k GPT model with the following configurations.

Step ii) Deploy model for test embedding.

Deploy model using same storage account.

In Azure AI Search, click on Import and vectorize data and select Azure Blob storage.

Select the storage account and the container where files are stored.

Select the text embedding model.

Create an index by specifying the name.

Create the indexer mentioning all required fields.

Select the GPT model from the Azure Open AI studio and select the data sources to add the grounding data from the Azure AI search.

Provide the data store details in the portal with vectorization using the text-embedding-ada model.

Select semantic search and API key for authentication.

Check whether the model can provide the answers based on the uploaded data.

Integrate the Azure AI Studio model with Copilot Studio

Connect the open AI GPT model to Copilot studio. Connect to the required environment. In the copilot studio, either create a new copilot or customize the created copilot. To add Azure Open AI data source, go to Generative AI-enabled topics and select Conversational Boosting. Click on Classic data and select Add connection on Open AI.

Set up the connection, as shown below.

Click on the connection and provide the endpoint and key to set up the connection. Provide the deployment name and other parameters.

Also, provide the index name and URL details. Select the index name based on the data source to be used.

Save the topic and test it. When any question is asked based on the data sources in Azure storage, Copilot provides answers and a link to the uploaded document.

The Copilot can be published on the required channels once it is ready.

Conclusion

Copilot Studio provides the capability to connect to Azure AI Studio to create customized Copilots. It also enables to connect to various data sources based on the storage used for uploading the data. The GPT models can also be selected based on the application scenarios. This enables complete customization based on the generative AI models as well.

About CloudThat

CloudThat is a leading provider of Cloud Training and Consulting services with a global presence in India, the USA, Asia, Europe, and Africa. Specializing in AWS, Microsoft Azure, GCP, VMware, Databricks, and more, the company serves mid-market and enterprise clients, offering comprehensive expertise in Cloud Migration, Data Platforms, DevOps, IoT, AI/ML, and more.

CloudThat is the first Indian Company to win the prestigious Microsoft Partner 2024 Award and is recognized as a top-tier partner with AWS and Microsoft, including the prestigious ‘Think Big’ partner award from AWS and the Microsoft Superstars FY 2023 award in Asia & India. Having trained 650k+ professionals in 500+ cloud certifications and completed 300+ consulting projects globally, CloudThat is an official AWS Advanced Consulting Partner, Microsoft Gold Partner, AWS Training Partner, AWS Migration Partner, AWS Data and Analytics Partner, AWS DevOps Competency Partner, AWS GenAI Competency Partner, Amazon QuickSight Service Delivery Partner, Amazon EKS Service Delivery Partner,  AWS Microsoft Workload Partners, Amazon EC2 Service Delivery Partner, Amazon ECS Service Delivery Partner, AWS Glue Service Delivery Partner, Amazon Redshift Service Delivery Partner, AWS Control Tower Service Delivery Partner, AWS WAF Service Delivery Partner and many more.

To get started, go through our Consultancy page and Managed Services Package, CloudThat’s offerings.