Microsoft wanted to make it easier for employees to answer what are often complex questions from customers about product licensing.
Licensing Navigator is an AI-driven agent built using Azure AI Services and Copilot Studio. The agent draws from hundreds of pages of content and was designed as a self-service resource for anyone in the company.
The agent answers over 1,000 questions a week from more than 800 monthly active users, enabling them to draft customer-ready responses almost 90% faster.
At Microsoft, licensing can be an especially complicated topic. For employees tasked with answering licensing questions—whether they’re sellers, support staff, business planners, or executives—the process of finding accurate information and interpreting nuanced policy language has often been slow and cumbersome. Even with well-crafted documentation and internal tools, the task of researching licensing policies could take an hour or more per question and still leave users uncertain about the accuracy of their responses.
Enter the Licensing Navigator, an AI-powered agent designed to transform how Microsoft employees access and interpret licensing information. Built using Azure AI Services and Copilot Studio, the agent was designed to make licensing guidance instantly accessible, accurate, and easy to validate.
“The Licensing Navigator agent has been a complete game changer for us in the way it leverages AI to identify and synthesize relevant information,” says Mark Bratton, Director, Monetization & Business Planning at Microsoft. “Employees no longer have to work through documentation and apply all that information to specific licensing scenarios. Licensing Navigator can do it all for them, instantaneously.”
“The Licensing Navigator agent has been a complete game changer for us in the way it leverages AI to identify and synthesize relevant information.”
Mark Bratton, Director, Monetization & Business Planning, Microsoft
The first version of Licensing Navigator was deployed within MSX, Microsoft’s internal CRM tool, primarily targeting sales roles that frequently handle licensing queries.
From the outset, the team made a deliberate decision to use only publicly available content, carefully curated to ensure relevance and accuracy. This approach was rooted in the long-term vision of making the tool customer-facing. However, working with such a vast and complex corpus of content posed significant challenges. To address this, the team employed a custom Retrieval-Augmented Generation (RAG) model using Azure AI Foundry, which allowed the agent to deliver precise answers by referencing semantically relevant chunks of documentation.
The process of preparing the content was meticulous. Nearly 600 artifacts were reviewed and reduced to only files with only the most relevant information. Each remaining piece was optimized for Azure AI Search-based consumption—headers were updated, tables restructured, and metadata tagged to enhance searchability.
Once the team was confident with the accuracy of the tool, Copilot Studio was used to orchestrate the various endpoints where the agent could be surfaced more broadly with employees.
The integration into Teams and M365 Copilot marked a major milestone, dramatically increasing accessibility and driving a surge in usage. Today, Licensing Navigator answers over 1,000 questions a week from more than 800 monthly active users.
“With Azure AI Services and Microsoft Copilot Studio, we had all the tools we needed to finely tune our search to power our agent on a hugely complex set of licensing documentation—and provide accurate responses in a user-friendly UI,” says Hunter Guffey, Senior Product Manager at Microsoft.
As the tool has evolved, the team has continued to improve the accuracy of the agent. “In fiscal year 2025, we more than doubled accuracy while expanding the content scope, thanks to ongoing engineering and content updates,” says Bratton. "We’ve shifted from simply making content more readable to optimizing the content for AI, and our model specifically, to help make sure it surfaces the right information.”
The team used multiple capabilities in AI Search and Azure OpenAI to improve performance. This included the semantic ranker that improves search relevance by reviewing the content and pulling the most semantically relevant chunks to the top. This process was combined with content tagging in Azure OpenAI and filtering in AI Search to better match content to questions. For example, a question intended for enterprise licensing but covered in content for other programs will only return answers from content tagged for enterprise.
Other improvements that have been especially popular with users is the agent’s ability to provide deep links to specific content sources. This not only boosts confidence in the responses but also streamlines the validation process. Multi-turn conversations were also added, enabling the agent to handle complex, context-aware interactions by maintaining conversational continuity across multiple exchanges.
A recent study demonstrated that access to the agent made drafting customer-ready answers to licensing questions (with supporting references) almost 90% faster than without the agent. Moreover, the overall quality of answers (a combination of accuracy and completeness) was 70% higher compared to using other AI tools and 140% higher than manual research.
Strong performance has led to equally strong user feedback. As one user said, “I use Licensing Navigator every day. The answer quality has gotten so much better, and I love the exact references it gives.” Another active user noted: “I just want to pour some praise onto the Licensing Navigator! I've tested it with several AHB (Azure Hybrid Benefit) questions, and despite this being a tricky topic, it gets them right. Whenever I get questions about AHB from now on, I'm going to suggest the questioner try the Licensing Navigator first.”
“With Azure AI Services and Microsoft Copilot Studio, we had all the tools we needed to finely tune our search to power our agent on a hugely complex set of licensing documentation—and provide accurate responses in a user-friendly UI.”
Hunter Guffey, Senior Product Manager, Microsoft
As Microsoft continues to refine the Licensing Navigator and expand its reach, the agent stands as a model for how AI can be used to simplify complex business processes and empower users at scale.
By harnessing the power of AI through Azure AI and Copilot Studio, the tool has turned a traditionally slow and complex process into a fast, accurate, and user-friendly experience. It empowers employees to self-serve, reduces the burden on specialists, and sets the stage for future innovations in AI-driven business support.
“Licensing Navigator is more than just a chatbot—it’s a paradigm shift in how Microsoft handles licensing support,” says Bratton.
Follow Microsoft