AI support for our IBM Support team
Cloud-based software has lots of components, and you can manage that when you have lots of agile development teams. So what happens if one of your many customers raises a problem with IBM Support - which team should get the ticket, so you can respond quickly?
Hursley develops and supports IBM App Connect, a key offering in IBM Cloud Pak for Integration. Thousands of businesses rely on it. We have dozens of development teams assigned to the product, using hundreds of Github repos.
Our Service Reliability Engineering (SRE) team needed a way to improve their manual process. Previously, they manually assigned a problem to a development squad, who may then direct it to another more appropriate squad after reviewing it. The objective: speed the problem through the system so the SRE team can respond to customers faster.
As the product has evolved, we've gathered data about fixes identified and the squads who sorted them. So now we can analyse that data to accelerate our identification of first responders.
This project uses IBM Watson Natural Language Classifier, which allows us to classify text into custom categories, at scale. The output helps our SRE team make better initial assignments when a new problem arises, from understanding information in an unstructured natural language description.
It also uses IBM App Connect itself, as an integration offering. It automates the classification, alerts the owning squad, and labels the issue as soon as a new support ticket is created. App Connect initiates a retraining of the model as more data becomes available.
Want to know more? No problem! Come and talk to us at the Festival.