How to Train Your Resources and Requests
You can solve tickets 5x faster and automatically resolve tickets by having strong training behind your Knowledge Base Resources (articles, links, and files) and Request Types. In this guide, we’re going to break down how to properly train both so that you and your team can start saving time!
The difference between a Knowlege Base Resource and a Request Type?
Knowledge Base Resources are great for questions like what is the “wifi password?” or “what’s our holiday schedule?”.
You can create a Knowledge Base Resource that answers these questions and in that way, you help your team find answers themselves. They also ensure that repetitive questions don’t make their way to your team, this helps you save a lot of time!
Request Type or service request helps you reduce the back and forth that happens when you’re trying to get information from the person making the request.
You can use a service request for something like “I need a new computer”, this request can prompt the person asking for a new computer with a set of questions like “what brand/model?” and “when do you need this?”. And on your end, you can assign tasks for yourself and others to complete
Training your resources and requests
Training your Knowledge Base Resources and Your Request Types will help you improve your auto-resolve rate and save time. the way to train both of these resources is very similar and can be broken down into 4 steps:
- Ask a question in the search bar
- Assign the request to the right team and to yourself
- Answer using a Knowledge Base Resource or update the Request Type
- Mark it as done
Let's take a look at how to train both separately.
Training Knowledge Base Resources
To train a Knowledge Base Resource, first, ask a question at the top search bar. then click file a request and you will be redirected to the request panel.
The key thing to always remember when training your resources is to assign the request to yourself and the right team. You are training machine learning here so every action makes the AI smarter!
Next, search for the resource that answers the question by clicking on the search icon. This will bring up a list of Knowledge Base Resources.
In this case, a resource already exists but if I need to create one on the fly, I would select the type of resource I want to create (import from Google Drive, Text, Link, or File) and add it as the answer. This is the best way to create Knowledge Base Resources.
Once I select the resource, “Requesting time off”, I add it to the request and mark it as “done”.
Now when I go back to the search bar and ask the same question, instead of prompting me to file a request, it shows me the correct resource:
By clicking “Yep”, I’m letting the AI know that this is the correct answer and it will increase the confidence score of this article. Your goal is to increase the confidence score of your Knowledge Base Resources, the higher the score the more trusted the resource is.
Currently, the confidence score is 3 –– the confidence score reads as “3 found this resource useful”. But the next time I ask that question, I will notice that the score has gone up to 4.
To increase the score, either click “Yep” where it asks “does this help?” or use the Knowledge Base Resource to answer a question on a request and then mark the request as “Done”.
Training Request Types (service requests)
Similar to Knowledge Base Resources, you can kick off the training by asking a question in the search bar at the top. In this example, we’re going to be making a request for a new computer. We’ve already created a Request Type for this sort of request called “New Computer”.
Incorrect Request Type returned
In the question bar, I wrote “my mac broke” and the Request returned isn’t the correct one, “New Computer”. Instead I was returned a Request Type called “Software or Network Problem”.
This is a perfect opportunity to train the model. To do this, change the Request Type by clicking the Request Type drop-down and selecting the correct one:
Once I do that, the questions will change. While training the model, you don’t need to answer these questions, you can file the request by clicking “Yes, file a new request”. You’ll then be redirected to the request panel.
Similar to how we trained the Knowledge Base Resources, first assign the request to yourself and then mark it as done.
The next time I make the same request, it returns the correct Request Type:
Expanding the natural language
With both Knowledge Base Resources and Request Type, it’s important to expand the natural language. We know that people ask for things in very different ways. For example, instead of someone saying “I need a new laptop” they may say “I need a new computer” or “I need a new PC”, that’s why it’s important to train your resources and requests using different wording.
With time, the more people ask for things in different ways, it will also expand the natural language but you can also get in front of this by asking for resources and services in different ways until you’ve trained your resources. To do that, we recommend pre-seeding your Knowledge Base with common questions.
By properly training your resources and requests, your teammates will be able to get the help they need faster. They’ll be able to self-serve with knowledge base resources or give you the information you need upfront to start working on their request.