Artificial intelligence (AI) and machine learning have become key components in helpdesk software. One of the biggest benefits of AI for helpdesks is that it can help eliminate repetitive administrative tasks. Ticket request routing and knowledge management are two categories of AI features that are part of helpdesks today. LBi’s HR HelpDesk is no exception.
AI is an important tool for the Employee Self-Service Portal of HR HelpDesk. It’s used for searching FAQs and for classification of newly entered Questions.
HR HelpDesk uses two machine learning models, one for each task:
- Suggesting an FAQ related to the user’s inquiry
- Classifying an inquiry into one of the existing categories
The learning process begins with data collection. The results produced for an FAQ Suggestion are only as good as the data that comes in. The model is “trained” on the text in the FAQ as soon as the FAQ is created to be able to offer suggestions to the user. Over time, data is gathered on the inquiries that were answered by the FAQ. This additional data allows the model to “learn”, improving the ability for the model to match different wordings of the same FAQ inquiry and to filter out irrelevant FAQs.
Service Level Agreements (SLAs) are the means for tracking and managing response times to resolve employee issues that are then measured against corporate commitment times (performance guarantees).
For instance, HR might guarantee a 24-hour (1 day) response to a paycheck or harassment issue, but up to 5 days to process a tuition reimbursement request. In many government regulated industries and unionized organizations, businesses may be required by law or contract to provide response guarantees, while other businesses may offer guarantees simply as a courtesy and for good will.
There is no better tool to manage SLAs than your HR HelpDesk — if you have one. Administrators set up broad case categories and specific subcategories within each category, then assign SLA time periods to each subcategory. From there, the system takes over and automatically tracks SLA performance in detailed reports.
So far, so good. But, how robust is the HelpDesk SLA configuration engine? Are your rules simple or complex? Do you measure SLA periods in hours or days? Do weekends count towards the SLA time? What about holidays or any other special days? Do they count? Are the rules different for different locations or employee classifications?