Microsoft Digital, the organization that is powering, protecting, and transforming Microsoft, is improving the support experience by partnering with ServiceNow to incorporate modern support-agent functionality into the Microsoft Digital environment by using ServiceNow Agent and Microsoft Teams. As a result, the support team, and the employees they assist have a more complete tool set, a simpler view into the support environment, and a more streamlined method for executing tasks and solving issues quickly.
Understanding virtual-agent-based support at Microsoft
Microsoft Digital runs the systems that support more than 135,000 employees. Our Global Helpdesk supplies support to these employees throughout more than 120 countries and regions worldwide. Global Helpdesk receives approximately 3,000 requests for support every day, and the ability to efficiently assess what help our users need and how we can provide that help are critical to the effectiveness of Global Helpdesk and our Employee Experience organization at Microsoft.
Examining the need for an improved agent experience
One of the primary touch-points for our employees has been agent-driven communication. Connecting to and speaking with a live support-team member has always been an important part of the end-user support experience at Microsoft. As part of our ongoing digital transformation, we’re working toward a more streamlined support experience for our support team and our users, including improved virtual-agent experiences.
The preexisting virtual-agent interface—one that Microsoft Digital developed and maintains—was initially developed as an experimental solution. While the previous solution had worked well in the past, we were experiencing several challenges in using an internally developed, custom-built tool:
Limited ability to innovate. While we initially developed the previous virtual-agent tool to meet our needs at the time of its implementation, our needs changed over the years. The way our employees work also changed, and industry changes and technology advances affected our platform’s efficiency. It was difficult to keep functionality and feature availability current on a self-developed platform.
Minimal integration and information reuse. The virtual-agent experience wasn’t effectively integrated with many of our other support tools, including our service-management platform knowledge base. Our support team had to copy and paste support information into the agent and manually search for and enter ticket management data for each support request.
Limited virtual-agent support. It was difficult to develop and maintain workflows for virtual-agent functionality and to provide the guidance and support that our live agents required.
Scalability and performance issues. Our previous platform didn’t support the scalability and performance goals that we set for our infrastructure and tools.
Creating a modern virtual-agent experience
Our vision for a modern support experience involves providing our employees with easy-to-use, transparent, and integrated services by transforming how we deliver support capabilities across all digital interactions at Microsoft. As part of this vision, we examined how we could improve the virtual-agent process while enabling our continued digital transformation in this area. We established several goals, including:
- Increase information reuse. Invest in integrated tools that support connected and correlated data and the reuse of information across support processes. We wanted a solution that would integrate directly with our service-management platform and integrate with existing data in our support environment.
- Take advantage of existing partnerships and tool sets. As part of our vision for simplicity and transparency in our infrastructure, we wanted to use existing tools effectively and pursue partners that supply complementary features.
- Adopt commercially available tools with out-of-the-box functionality. We wanted to move to a commercially available product that was already in development for a large customer base. We wanted a solution that enabled agile development and also provided built-in support.
- Identify AI and machine-learning capabilities to streamline and improve support and employee experiences. Virtual-agent functionality was an important consideration for the new solution—whatever we chose needed to integrate AI and machine-learning tools to better inform the support process and better serve our Helpdesk users.
Combining ServiceNow and Teams
To begin the process of selecting a new platform, we used our challenges and goals to evaluate several potential solutions. The ServiceNow Agent suite of capabilities now serves as our primary agent experience for live and virtual-agent support, and we use Teams as one of our most critical agent hosting environments. We’ve used ServiceNow IT Service Management for six years as our primary support-automation tool. Microsoft and ServiceNow have engaged in a strategic partnership to accelerate digital transformation for enterprise customers. As part of the partnership, we’re working together toward a common solution based on ServiceNow and Microsoft platforms, with a goal of building better products and improving customer experience for both companies.
By combining ServiceNow and Teams, we’ve created an agile support platform that easily integrates with our existing support environment and data. We’re using all the ServiceNow out-of-the-box features that fit our needs to ensure that our use-case scenarios are robustly supported. Our current ServiceNow implementation included several important milestones and challenges. Highlights from the implementation process include:
Aligning feature capabilities and requirements. We prioritized features and implementation timelines, noting what was necessary to decommission the previous solution and facilitate a smooth transition.
Establishing a crawl, walk, run approach. Our implementation started small, and we iterated from there. We incorporated quick-win features early, ensuring that we established sound implementation practices before moving on to larger and more important components.
Identifying and resolving Azure AD synchronization issues. Initially, Azure Active Directory (Azure AD) synchronization complexities made it difficult to fully integrate user data. However, through our partnership with ServiceNow and because of ServiceNow’s agility and large product support team, we received an update in the next release that fixed the problem.
Supplying direct agent integration across multiple contexts. While Teams is one of our primary agent interfaces, we’re also incorporating the ServiceNow agent across many of our web-portal workspaces, offering in-context agent support to our employees within the application or interface that they’re using.
Integrating with Azure Cognitive Services components for intelligent virtual-agent functionality. We’re using the Azure QnA Maker service to supply natural-language FAQ support to the virtual-agent experience, using existing knowledge-base sources.
Results and benefits
Our new support-management platform has generated several benefits, for our business, our support agents, and our customers. These benefits include:
- Simplified interaction for our end-users. Our virtual-agent usage has doubled during the first year using ServiceNow. More of our Global Helpdesk end-users are using the virtual agent to initiate support contact. Almost 50 percent of virtual agent-initiated issues are solved within the virtual agent context, without needing human involvement. The increase in virtual agent adoption decreases support costs and allows our live agents to focus on more complex issues. We expect virtual agent adoption and efficiency to increase as we improve functionality, build the virtual-agent interface into more of our end-user environment, and grow virtual-agent adoption as a replacement for other contact methods, such as email or phone calls.
- Increased future cost savings from the new solution platform. By implementing a software as a service (SaaS) solution like ServiceNow Agent, we benefit from the built-in resiliency, scalability, reusability, and feature set of a SaaS platform. We’ve also deprecated our internally developed tool, saving the associated infrastructure and maintenance costs.
- More effective self-service problem solving for our end-users. The ServiceNow virtual agent enables detailed workflow automation within the agent interface. We’re using that automation to create a better self-service experience that allows our users to resolve issues without live-agent intervention, giving more time back to the live agent.
- Improved live agent handoff. We’ve significantly improved integration between the virtual agent and our other support systems data. When the virtual agent transfers an end-user to one of our live agents, the live agent can access the virtual-agent chat transcript alongside the ServiceNow ticket and end-user information. This capability makes it easier to provide comprehensive support for the end-user in a live chat experience.
- More accurate metrics for end-to-end support processes. Due to integration with ServiceNow and other data platform, we can better understand the end-to-end support experience and identify opportunities for improvement and greater efficiency across the entire support landscape. This enables us to further reduce support costs and decrease support resolution times for our end-users.
We learned several important lessons throughout the implementation of the ServiceNow Agent with Teams at Global Helpdesk, including:
- Monitor the integration process carefully. Accurate monitoring was crucial to the success of our integration. By identifying issues with data integration, identity synchronization and general functionality, we were able to catch potential problems early in the integration process before they negatively impacted the end-user experience.
- Identify expected failure scenarios and prepare your team to address them. Unanticipated failures take more time and effort to identify and fix. Monitoring helped us understand the likelihood of certain failure scenarios, which we prepared for by using knowledge articles and performance dashboards.
- Identify where the natural-language understanding needs training. We found that effective training for natural-language understanding was critical to virtual-agent success. We tested and refined our training process for these models until we achieved an accurate and effective model that supplied optimal natural-language interactions for our virtual-agent users.
Implementing the ServiceNow Agent and Teams solution has created several key benefits already. One of the most significant benefits is the ability to quickly iterate and identify new features and improved functionality for the near future. We’re currently performing a post-implementation review to ensure that we incorporate our lessons learned into future platform refinements, and we’re working on several improvements:
- Integrated monitoring. We’re running a proof of concept to integrate monitoring into the virtual-agent experience on Teams and demonstrate how monitoring data integration into the agent experience can create a more proactive support environment.
- Sentiment analysis. We’re refining sentiment-analysis capabilities to determine the overall sentiment of users’ interactions with the agent, whether those sentiments are positive, negative, or somewhere in between. These sentiment scores help us evaluate overall satisfaction and ensure that agent interactions that require immediate attention are flagged immediately.
- Bot-to-bot integration. We’re improving the virtual-agent-to-live agent handoff process that will seamlessly transfer a support user to a live agent if the user’s issue isn’t resolved in the virtual-agent experience. This handoff involves all recent and historical interactions, a record of the virtual-agent workflows that have been triggered, and any other valid data obtained from the virtual-agent experience.
Integrating ServiceNow Agent and Teams is moving us toward a single interface in the support experience, a complete tool set, a simpler view into the support environment, and a more streamlined method for executing support tasks. We're using machine learning and predictive analytics to enable our agents to access all the information that they need to resolve an issue—including information that they might not know they need. We're using existing ticket data to supply insights into the resolution process so that our agents can supply the quickest resolution possible. We’re implementing intelligent workflows that are automatically presented to the agent based on issue classification to help guide the agent through the troubleshooting process with the user. As a result, our support team is better equipped, our employees are better supported, and our business runs more smoothly.
© 2021 Microsoft Corporation. This document is for informational purposes only. MICROSOFT MAKES NO WARRANTIES, EXPRESS OR IMPLIED, IN THIS SUMMARY. The names of actual companies and products mentioned herein may be the trademarks of their respective owners.