Addressing Customer Inquiries: Choosing Between Copilot Studio and Azure AI Foundry for Agent Development
Choosing the Right AI Tools for Employee Support
Introduction to Microsoft AI Solutions
When organizations seek solutions for their internal challenges, selecting the right technology is critical. Recently, a customer inquired whether to use Microsoft Copilot Studio or Microsoft Azure AI Foundry for creating an employee self-service agent. Inspired by this question, the team at Microsoft Digital, led by Ravi Goriparthy, a principal software engineering manager, embarked on a project aimed at building a more efficient support mechanism for their Human Resources (HR) and IT departments.
The Goal of the Project
The objective was to develop an intelligent agent that could:
- Answer inquiries related to HR policies
- Provide IT assistance
- Interact with various Line of Business (LoB) applications
This ambitious project faced the challenge of selecting the right combination of Microsoft tools—specifically, Microsoft 365 Copilot, Microsoft Copilot Studio, and Microsoft Azure AI Foundry.
Navigating the AI Landscape
Goriparthy and his engineering team began by assessing their existing LoB applications, focusing on policy information, help content, and employee data. Goriparthy noted the importance of finding an agent capable of smooth integration with these diverse applications, particularly since much of their data was located outside the Microsoft 365 ecosystem.
“We needed an agent that could easily integrate with a lot of LoB apps,” Goriparthy emphasized. Their investigation pointed them towards Copilot Studio, which is part of the Power Platform and boasts connectors for over 1,000 LoB applications.
Building the Self-Service Agent
Using Copilot Studio, the team developed an agent that incorporated various SharePoint sites as knowledge sources. However, they encountered a hurdle where responses were sometimes retrieved from unmanaged SharePoint sites.
To mitigate this issue, they collaborated with the Copilot Studio and Microsoft 365 Copilot teams to ensure answers from authorized SharePoint sites were clearly marked with an "authoritative source" badge.
Fostering Integration and Enhancements
The project did not stop at initial development. The team also integrated Copilot Studio with SuccessFactors and ServiceNow, identifying and addressing gaps in the existing Power Platform connectors. This effort included adding a metadata-driven extensibility design, allowing custom agents to dynamically modify API calls for better data management.
“These integrations opened up numerous opportunities to improve employee productivity by providing an agent that can read employee data and take action on it,” Goriparthy noted.
The agent not only enhanced the experience of Microsoft employees but also showcased how adaptive technology can elevate workplace productivity and engagement.
Effective Use Cases of Microsoft AI Solutions
Microsoft’s suite of AI tools can be categorized based on their primary use cases. Here’s a brief overview of when to use each tool:
Microsoft 365 Copilot
Designed to enhance productivity within apps like Outlook and Teams, common use cases include:
- Document Creation: Assisting in drafting documents, presentations, and data analysis.
- Email Management: Summarizing emails and drafting responses.
- Meeting Management: Generating summaries and action items.
- Project Management: Streamlining planning and tracking.
- Collaboration: Enhancing teamwork through integration in SharePoint and OneDrive.
Microsoft Copilot Studio
As a low-code tool for custom AI agent creation, it’s suitable for:
- Customer Support: Automating responses to customer queries.
- Employee Assistance: Managing inquiries related to HR and employee benefits.
- Custom Solutions: Tailoring Microsoft 365 Copilot for specific business needs.
- Sales and Marketing: Supporting sales inquiries and facilitating marketing efforts.
Microsoft Azure AI Studio
Ideal for more specialized AI applications, use cases here include:
- Generative AI: Creating applications that produce content and insights.
- Predictive Analytics: Developing models to forecast trends.
- Natural Language Processing: Analyzing text data for insights.
- Image and Video Analysis: Enhancing processes like quality control.
- Custom AI Solutions: Addressing specific needs across diverse industries.
Throughout this project, Goriparthy encouraged others to share their experiences and insights regarding AI, offering guidance on which tools to deploy depending on organizational needs.
Getting Started with AI Agents
For organizations looking to implement agents or Copilot Studio, here are some practical steps:
- Evaluate Needs: Assess your organization’s requirements to determine the suitable tools.
- Configure Knowledge Sources: Identify relevant knowledge sources and ensure they are appropriately labeled.
- Enhance Connectors: Build custom connectors where existing Power Platform connectors may be lacking.
- Boost Productivity: Utilize these integrated agents to streamline processes and improve overall productivity.
- Share Insights: Collaboration and communication of experiences can help others in navigating the AI landscape effectively.
Organizations exploring AI solutions can benefit from proactive engagement with these tools, paving the way for enhanced productivity and streamlined processes.
