Microsoft's internal IT organization, Microsoft Digital, has deployed an AI system that automatically keeps support knowledge bases current, projecting 16,000 hours saved annually and a 10% reduction in support tickets. The AI for Knowledge Management platform mines support ticket data to identify content gaps and generate article updates, cutting manual review cycles that once required a dedicated team's full-time effort.
Fragmented knowledge and manual reviews
For the Global Help Desk, the problem wasn't just volume. Support knowledge lived across thousands of self-service articles, agent-facing systems, and SharePoint sites, all changing constantly. "If the knowledge isn't accurate and current, the experience breaks down immediately," said Kevin Verdeck, a senior IT service manager in Microsoft Digital. "Bad content leads to bad answers."
Keeping that content current meant a manual review process. Teams analyzed usage data, relied on support agents to flag missing or outdated material, and ran recurring review cycles with subject-matter experts. One five-person group reviewed 1,900 self-service articles and 1,700 agent-facing articles every six months-and that didn't include SharePoint sites. "It was basically their full-time job doing regular reviews," Verdeck said.
When employees couldn't find answers, issues escalated to advanced support. The system was reactive and hard to scale, and knowledge gaps often surfaced only after someone hit a dead end.
Turning raw data into knowledge
Microsoft Digital built a centralized AI pipeline that turns every resolved support ticket into a signal. The system ingests large volumes of incident data, cleans and structures noisy ticket details, and clusters patterns to spot recurring issues. "When you have a large volume of data, it's a silent gold mine," said Ankit Guddewala, a software engineer on the team. "The sheer brilliance lies in taking that data, making it sing, and letting it tell you exactly where the treasure is."
The pipeline then compares those patterns against existing knowledge articles. If the match is below 40%, it flags a need to create new content. Between 40% and 80%, it suggests updating existing articles. Above 80%, no change is needed. Subject-matter experts receive email notifications to validate the recommendations, keeping a human in the loop. Their feedback also tunes the AI model over time.
"Our goal is to free people from the manual work of maintaining content so they can focus on improving its quality," said Namrata Ladda, a product manager in Microsoft Digital. "With the right human-in-the-loop balance, AI can do the heavy lifting while people make sure the final knowledge is accurate and useful." This approach aligns with broader shifts in AI for Customer Support, where content accuracy directly determines whether self-help channels succeed.
Impacts and what's next
The Global Help Desk now identifies knowledge gaps without waiting for reports. AI generates structured article drafts, so humans can focus on quality rather than hunting through data. The team estimates the solution will save 16,000 hours annually, reduce support tickets by 10%, and lower the number of advanced support escalations.
"Turning our knowledge base from a static thing into a living knowledge base is a big step forward," Ladda said. Other Microsoft teams, including HR, have shown interest. Once internal testing completes, AI for Knowledge Management will be released to all company employees and customers.
"Knowledge management today is about making sure people can find the right answers the moment they need them," said Silvina Olkies, senior director of Service Management in Microsoft Digital. "When knowledge stays current, employees get unblocked faster and productivity improves, making the overall support experience far more efficient."
Why this matters for management
This project shows how AI for Management can target a specific operational bottleneck-knowledge maintenance-and deliver measurable time savings. For managers, the takeaway is concrete: audit how your team keeps support content current, identify the manual steps that consume the most hours, and apply AI to those steps first. The Microsoft Digital team didn't try to automate everything at once; they started with the signal already buried in ticket data and built a pipeline around it.
Actions your organization can take now:
- Treat your knowledge base as the source of truth that determines whether AI and self-help succeed.
- Audit how knowledge is maintained, looking beyond publishing workflows to see how gaps show up in real usage.
- Identify where people spend the most time searching and reporting knowledge, and target those steps for automation.
- Use AI to maintain content, not just serve it-let it flag gaps and refresh out-of-date information.
Your membership also unlocks:
