Extending Copilot Studio with SharePoint Search: Overcoming Conversation Boosting SharePoint DataSource Limitations

In the ever-evolving landscape of digital collaboration, Microsoft Copilot Studio has emerged as a powerful tool for enhancing productivity and streamlining chatbots. However, like any technology, it has its limitations. One such limitation is the challenge of effectively boosting conversations within SharePoint. This blog post explores how integrating the SharePoint Search engine can overcome these limitations, providing a more robust and efficient solution.

Understanding the Challenge

Copilot Studio is designed to assist users by providing relevant information and suggestions during conversations. However, when it comes to SharePoint, the vast amount of data stored in various formats and locations, as well the current LLM model (ChatGPT 3.5) can make it difficult for Copilot to retrieve and present the most pertinent information. This can lead to inefficiencies and missed opportunities for getting the right results.

Key Limitations of SharePoint Data Source for Copilot Studio:

As stated on Use SharePoint or OneDrive content for generative answers - Microsoft Copilot Studio | Microsoft Learn, due to memory limits, generative answers can only use SharePoint files that are under 3 MB. If a file is larger than 3 MB, consider splitting it into multiple smaller files to ensure that Copilot Studio can effectively process and utilize the content. Although I understand this requirement, is not feasible that an enterprise with a big Intranet or a big sized document management is able to do this.

So this is where the SharePoint Search engine can come in handy.

The Power of SharePoint Search

SharePoint Search is a powerful tool that indexes content across SharePoint sites, making it easier to find relevant information quickly. By leveraging the capabilities of SharePoint Search, we can enhance Copilot Studio’s ability to retrieve and present information, thereby overcoming the limitations of conversation boosting within SharePoint.

Integrating SharePoint Search engine with Copilot Studio

1. Indexing Content: The first step in integrating SharePoint Search with Copilot Studio is to ensure that all relevant content is indexed. This includes documents, lists, libraries, and other data stored within SharePoint. Proper indexing ensures that the search engine can quickly locate and retrieve the necessary information.
2. Custom Search Queries: By creating custom search queries tailored to the specific needs of your organization, you can refine the search results to be more relevant and useful. These queries can be designed to prioritize certain types of content or to filter out irrelevant information, making it easier for Copilot to provide accurate suggestions.
3. Enhanced Metadata: Adding metadata to your SharePoint content can significantly improve search results. Metadata provides additional context about the content, making it easier for the search engine to understand and categorize the information. This, in turn, helps to retrieve more relevant data during conversations.
4. Search Refiners: Utilizing search refiners allows users to narrow down search results based on specific criteria. This can be particularly useful in large organizations where the volume of data can be overwhelming.
5. Continuous Improvement: The integration of SharePoint Search with Copilot Studio should be an ongoing process. Regularly reviewing and updating search queries, metadata, and indexing practices ensures that the system remains efficient and effective. Additionally, gathering feedback from users can help identify areas for improvement and guide future enhancements

How to achieve this?

First of all, you need to go the conversation boosting topic and create a Power Automate action:

On the input variable, you are going to send the "Activity.Text" variable.

Now, you need to create the flow. The first action you need to add is the "Send an HTTP request to SharePoint". Here is the trick, you can now use the SharePoint API to make queries to SharePoint engine. A good tool that I would recommend for you to test these queries is the SharePoint Search Query tool. This could be a good way to enhance the results that you need.

Next, obviously you need to use the Parse JSON action to parse the results.

After that, you need to create the array of the search results.

Now, we will map the results to a format that CoPilot understand on the "Create generative answers" - you can find the necessary format here -> Use a custom data source for generative answers - Microsoft Copilot Studio | Microsoft Learn.

Finally, you need to return the results in a string format to CoPilot Studio.

Your complete flow should look like this:

Now going back to the CoPilot Studio, you need to parse the results to a table, create a new condition in case that has found results.

Now, you can use the "Create generative answers" with Custom data source with the following settings:

• Search only selected sources
• On custom data, select the table variable
• Content Moderation should be "Low"

Looking Ahead

The integration of SharePoint with Copilot Studio is just the beginning. Microsoft is committed to continuously improving our tools to meet the evolving needs of our users. Future updates will bring even more enhancements, further empowering you to achieve your goals with greater efficiency and effectiveness, so remember this is just a workaround and as any workaround as it's own limitations, being one of the them the need for enrichment the data for the custom data.

Feel free to share your thoughts and experiences with this new extension in the comments below. I’d love to hear how it’s making a difference in your daily copilots!

#CopilotStudio #SharePointSearch #DigitalTransformation #Productivity #Innovation #TechUpdates