Using AI to Manage IT Data
Just like any other organization, MSPs require accurate data to keep the business moving forward and growing. But how can you improve your operations efficiency when your data doesn’t properly identify the problems? It’s an example of the classic garbage in, garbage out. The goal is to find ways to clean up the mess!
One of the [many!] issues our MSP faced was accurate ticket categorization. Ticket categorization is important to determine widespread issues, or recurring support issues from a particular client or with a particular piece of software. Categorization helps us identify these incidents and address them quickly. ConnectWise, our ticketing system, mitigates this via “Type, Subtype and Item” categories that can be associated with every ticket. A Type is a broad categorization of service support categories. A Subtype is a more specific classification within a Type category. An Item is a specific entry within a Subtype category. These terms are used to organize and classify tickets based on the nature of each IT support request.
Sound straightforward and easy? Not so much. The two major issues with Type, Subtype and Item classification are:
- Creating the Type, Subtype and Item categories and maintaining them in an intuitive structure.
- Ensuring that the engineers properly set each category after completing a support ticket.
After two unsuccessful rounds of developing the categories and attempting to maintain an intuitive substructure, we still had “garbage in.” We weren’t getting the correct search results. In an attempt to avoid a strikeout on round three, we decided to look at AI based alternatives and settled on CrushBank.
Setting up CrushBank was a breeze – just a new API key and the ingestion of two years’ worth of previous tickets. After setup, the first module we enabled was SmartClassifier to address our Type, Subtype and Item issue. SmartClassifier can ingest and learn your current Type, Subtype and Item if you want to keep your custom structure intact. It can also be pre-loaded with its own default categories that cover the majority of support issues, with an option to add additional Type, Subtype and Items if you want to start from scratch. As new tickets are created, SmartClassifier will use AI to review the ticket and automatically assign the Type, Subtype and Item, saving engineer time.
The Verdict on AI is In
Overall, our experience has been positive with SmartClassifier. It certainly saves time on the engineer side and reduces the overhead of category management. The AI is constantly evolving and “learns” over time to improve accuracy. This is especially useful with the custom categories that CrushBank was not preloaded with or “aware” of.
But wait, there’s more! Two other CrushBank components we recently implemented are Ticket Budgeter and Auto Summarization. Ticket Budgeter works exactly as it sounds and adds an estimated budget of hours to complete a given ticket. The AI reviews historical tickets with similar scope and assigns the hours accordingly. On the other hand, Auto Summarization crafts concise one- to two-line ticket summaries based on the intricacies of the support request. We’re still kicking the tires on both modules, but initial feedback has been overwhelmingly positive.
As we dive into the MSP world of support, it’s obvious that AI is the new kid on the block and ready to shake things up. CrushBank is like our cool sidekick, giving us a taste of what AI can do for MSPs. This is just the start of the AI adventure. CrushBank is the opening act to a wild ride into the unknown, where AI is the rockstar changing the game for managed service providers. Get ready for the rollercoaster because we’re just getting started!