Copilot Chat to Dataverse Storage
Budget: $250 – $750 AUD
I’m building a text-based Copilot experience inside Microsoft Teams (within our M365 tenant) and need a developer who can take the conversation a user has with the bot and store each session's submission as a new row in a Dataverse table.
The chat is not scripted— prompts change on the fly— but there are standardised questions that give the conversation some structure. So the bot must generate dynamic follow-ups and acknowledgements, and when confirmed by the user, log the responses in Dataverse. The table data structure has been defined with minimum required fields and some recommended or helpful fields, but can be modified based on your suggestions.
What I already have:
• A licensed M365 environment with Teams and Dataverse enabled.
• High-level sample user stories for the conversation flow.
• A working prototype of the agent that needs improvement
What I need from you:
• A fully functional Copilot/Teams app that initiates and maintains the chat, calls the AI service for dynamic responses, and writes each user message to a specified Dataverse table.
• The updated Dataverse schema and any necessary environment variables or connection references.
• Deployment package and step-by-step instructions so I can import, test, and publish the bot in my tenant.
• The ability to view the contents of the Dataverse table in a web-based view or something similar.
• Brief documentation outlining how to extend the table or change prompts in the future.
Acceptance criteria
1. User can add the Copilot to any Teams chat or channel and start a conversation.
2. Every user submission appears as a new row in Dataverse within two seconds, with no data loss during long sessions.
3. Source code and configuration files compile and deploy without modification inside a standard M365/Dataverse environment.
I can try to grant you access to my copilot studio environment to improve my idea, but if access doesn't work, you must develop the agent in your copilot studio environment and power apps etc, and deploy it to your MS Teams to prove that it can be accessed via teams. once im happy with it, you will then have to provide me with the agent config and dataverse schema so i can replicate on my end easily.
The chat is not scripted— prompts change on the fly— but there are standardised questions that give the conversation some structure. So the bot must generate dynamic follow-ups and acknowledgements, and when confirmed by the user, log the responses in Dataverse. The table data structure has been defined with minimum required fields and some recommended or helpful fields, but can be modified based on your suggestions.
What I already have:
• A licensed M365 environment with Teams and Dataverse enabled.
• High-level sample user stories for the conversation flow.
• A working prototype of the agent that needs improvement
What I need from you:
• A fully functional Copilot/Teams app that initiates and maintains the chat, calls the AI service for dynamic responses, and writes each user message to a specified Dataverse table.
• The updated Dataverse schema and any necessary environment variables or connection references.
• Deployment package and step-by-step instructions so I can import, test, and publish the bot in my tenant.
• The ability to view the contents of the Dataverse table in a web-based view or something similar.
• Brief documentation outlining how to extend the table or change prompts in the future.
Acceptance criteria
1. User can add the Copilot to any Teams chat or channel and start a conversation.
2. Every user submission appears as a new row in Dataverse within two seconds, with no data loss during long sessions.
3. Source code and configuration files compile and deploy without modification inside a standard M365/Dataverse environment.
I can try to grant you access to my copilot studio environment to improve my idea, but if access doesn't work, you must develop the agent in your copilot studio environment and power apps etc, and deploy it to your MS Teams to prove that it can be accessed via teams. once im happy with it, you will then have to provide me with the agent config and dataverse schema so i can replicate on my end easily.