VMware Customer Connect delivers continuous, agile releases to improve our customers’ experiences. I’m excited to share a new Customer Connect AI capability that uses Machine Learning (ML) technologies to proactively suggest resolutions or troubleshooting steps to solve technical issues. This new capability offers tremendous value to our customers. All our releases are designed to help customers self-resolve issues and find helpful resources faster, but this one is particularly exciting. Let’s walk through how this works using an example.
How It Works at a High Level
Let’s say a customer is having trouble solving a certain technical issue and needs to file a support ticket. They go to the technical support request filing page on VMware Customer Connect where they select the product and describe the problem. This is where some amazing intelligence technology comes into play. Customer Connect AI runs the issue description through two types of ML technologies, Watson Assistant and Natural Language Classifier (NLC). The reason we’re using both technologies is to ensure space-travel-precision accuracy of the right solutions (well maybe not quite that level, but we did spend a lot of time trying to make this thing accurate) of when to present the help. We want to avoid presenting troubleshooting steps for the wrong issue – that’s customer service 101! We set up an accuracy percentage threshold that must be met before Customer Connect AI suggests the resolution based on the issue description.
If the customer cannot resolve their issue via Customer Connect AI’s troubleshooting steps, then they have the option to chat with a live Technical Support Engineer (TSE) for further investigation or to file a support request online. All attempted troubleshooting steps will be shared with the TSE in the case ticket whether the customer chats with them or files a ticket online. This is a valuable feature because no customer wants to have to repeat things they have already tried when they engage with a TSE. Besides, customers just want to know “I did a, b, c, d, e on my own and none of them worked, now tell me what else I should try that’s different.” Our goal is to provide issue resolution in the fastest time possible and this cool new feature is paving the way for even more innovative opportunities. Watch the brief demo below to get a feel for how this new feature works.
Starting Small with Plans to Expand
We’re rolling out this feature with an initial sampling of three use cases targeted around vCenter and ESXi products in this release. The problem categories these three use cases fall under include: vCenter Server Management, Infrastructure Service Management (PSC/SSL), Installation / Upgrades. I know, three is a small number, but we identified use cases with high-ticket volume and problem resolutions which are relatively straightforward. Because this is the initial release, we’re starting small to prove the concept.
Our goal in the coming quarters is to scale the number of use cases exponentially with other products and VMware solutions. We’re also exploring embedding more ML and Natural Language Processing into the actual troubleshooting steps, so they become even smarter and more dynamic.
I hope this new functionality will help you resolve your issues faster. Feel free to share your questions and/or comments.
