IBM has announced that it is integrating Watson to service desks. It is offering the cognitive computing solution as a service to clients called IBM’s Workplace Support Service with Watson. Its intention it to bring the power of cognitive computing to handle the majority of support tickets once it is integrated into other automated helpdesk functionality.
Supporting software roll outs
This does not mean that Watson will fully replace human interaction but will gradually learn resolution to problems over time. Those it is unable to handle will be passed seamlessly onto a human service agent. The example IBM gives is an Office 365 roll out. Watson has already ingested all the Microsoft knowledge base articles including the related ones. It helps a typical user through the installation, offering guidance based on those articles when issues arrive. With its natural language interface Watson is able to deliver a service that frees up support teams from the majority of support calls.
Richard Esposito, general manager, GTS Mobility Services, IBM commented: “Today, governments and enterprises need to provide an effective set of capabilities to their workforce, so that their employees can deliver a superior interaction and experience for their citizens and consumers. We need a system that can understand and communicate in a natural language conversation, one that solves problems and continues to learn while engaging with employees. Our Workplace Support Services with Watson delivers this value.”
Follow-the-sun becomes the norm
For organisations looking to provide support services around the world the Watson solution can deliver significant benefits. The cognitive computing platform can not only ingest support articles that assist its capabilities, it can also learn from its own interactions. This allows the platform to support individuals using any device, in any location at any time. Watson will also deliver a personalised experienced based on an individual’s requirements. This personalisation means that as Watson it responds to individuals based on its knowledge of their technical expertise.
Watson is also able to scale rapidly to meet a spike in requirements from deployments or changes across the globe. As new information becomes available about resolutions Watson can be taught immediately avoiding delays in a traditional service desk deployment. Watson supports more than 20 languages with localized display support in English, Japanese and Spanish. That list will grow over time and it is always improving its grasp of syntax and vocabulary. It understands and is constantly improving its knowledge of cultural differences and dialects.
Prescriptive analytics with added analytics
Watson can also deliver on the promise of prescriptive analytics. As it analyses information it receives, it is able to identify actions that need to be taken to resolve the issue. This does not always mean that those steps are undertaken by the caller. Watson can initiate the actions itself to resolve the call. For example where a user has run out of storage it can extend storage within an Exchange environment or reset a password. Wherever an action is identified as a resolution and it is possible to automate the process, Watson is capable of supporting it. This could be as diverse as ordering a new phone for a user creating a field service appointment to repair hardware.
Watson learns from its interactions. As it gathers information about the support calls it answers it narrows down the optimal paths for resolution. This not only reduces the time take for callers to have their issue resolved but also frees up Watson to scale, handling even more calls per hour. It engages with users in a natural language interface that means when users are transferred to a human agent the change of control for the call is seamless.
IBM’s Workplace Support Service with Watson will transform how service and support is delivered. It is integrating IBM’s helpdesk services with analytics and Watson’s cognitive capabilities. This allows it to learn from user behaviour and improve over time. It can support individuals on any device, any time and at any location. It will offer faster resolution of IT issues, handling the majority of support tickets when integrated with other helpdesk automation functions.
This is not the first solution to use machine learning algorithms to create support tickets. Knoa Software is doing a similar job with its ERP solution that Boticario in Brazil has taken on.
How far will Watson bots go?
Conversations between users and helpdesks are probably the most fractious internal communication for any organisation. This raises the question as to how far IBM will develop Watson’s capabilities. For example, will it use the Watson Emotion APIs to detect how upset a user is? This would allow a call to be routed to a person rather than the bot if the caller was upset. If the caller is angry, maybe the call is handled by the bot to reduce the risk of damaging interpersonal communications.
Watson also has Tone APIs that some customers are using to detect if callers are telling the truth. This is very useful for fraud teams and equally useful for support teams. Having managed support desks there are times when there is little trust in what the customer is saying.
The more important benefit here is in being able to move beyond the obvious problem. In a help desk situation the problem can often be more than just the software the user is calling about. Watson would be able to apply a much wider set of knowledge and analytics to the call and identify other areas for investigation. This is often where experienced support technicians rely on gut instinct without being able to explain how they got to the solution. Watson would, at least, be able to explain its reasoning.
This announcement was long expected. Where Watson will see the most traction for this solution is difficult to say. The initial inference is software support. Telecommunication companies and banks will also consider the technology for their support services. The recent partnership with Salesforce could see it supplement many Salesforce Service Cloud installations as well. Once the integration API’s are completed it will be interesting to see what customer announcements are made.
Whether this will see the end of call centres is too early to say. For some companies looking to make significant investments it is an option they will seriously consider in the future. This could be a huge revenue stream for IBM and Watson in the future. It will be interesting to see how quickly it is taken up and how quickly IBM can scale it. What will be interesting is whether Google DeepMind follows IBM into this space. If IBM proves there is a significant market here then Google will follow, as they did in the healthcare market.