Learn how to automate requests
Welcome to Spoke! Follow along to learn how to automate both knowledge requests and service requests. If you already know how to automate knowledge requests, you can skip to the Service Request portion of the article using the contents on the left.
Automating Knowledge requests
Type your request into the request bar at the top of Spoke. Press Enter to submit your question.
Because your knowledge base is (mostly) empty, Spoke will ask if you'd like to create a request. Select Yes. This will take you into the expanded request view.
The expanded request view is the central place where Admins and team members can respond to requests and create knowledge.
To add knowledge to the knowledge base, click the + button
- Beside Write a reply", click on the + button.
- Choose to add a Google Drive item, File, Link, or Text. Click Search KB to send a resource that already exists in the knowledge base.
- Complete the answer and make sure to add keywords.
- Turn on Automate this reply. Since you asked the question Spoke assumes you don't know the answer, so this is automatically switched off.
- Click Add.
- On your keyboard, press Enter to send the answer to the end requester.
- Click mark as done on the upper right to resolve the request.
Reinforce the connection between question and answer by:
- Asking similar questions in natural language from another channel.
- Adding the same answer again by clicking Search KB instead of Google Drive, File, Link, or Text.
- Make sure you're telling Spoke what articles are and are not helpful.
Automating Service requests
When a user needs help that can't be solved by knowledge, such as hardware to be fixed, or an address to be updated in the HRIS, there's additional information needed to process the request. Chasing down a user to find answers is never fun, and often time consuming.
Request types allow you to automate requests in two ways:
- Ask questions of a user at the time the request is filed
- Automatically create tasks to better work cross-functionally
- Navigate to "Teams" > Choose the team for which you're building a request type
- Select the "Request types" menu.
- Select the red "+" in the bottom right of the screen. Choose an existing template to customize here.
Choose a template from a different team by clicking the downward arrow in the top left of the modal, and choosing a new team.
Choose a template from the list. In this case, we chose "New hardware."
Update the title and description of your request, then choose an icon by selecting the downward arrow in the right corner of the icon box.
Select Questions to add or edit questions
To add more questions, Select "Add a question"
Choose the field type required:
- Text (Freeform response) - This allows your users to write short responses to the question. (This field has a limit of 100 characters)
- Dropdown form (select one) - This allows you to create a dropdown and a list of choices from which to choose.
Once you've selected the Request Type you'd like to edit, navigate to the "Tasks" tab.
Choose "+ Add a task" to enter and assign a task.
Enter the task that needs to be completed, then select the downward arrow next to "leave it unassigned" to determine the task assignment.
Here, you will be given a list of options to assign this task.
- Request Assignee - The person handling the request has specific action items each time this request is created.
- Requester - The person filing the request needs to reply with information such as a signed handbook or tax documents.
- Requester's Manager - If a user's manager needs to approve items such as hardware or software purchases.
- A Member of... - any member of the chosen team.
- A Specific user... - a certain subject matter expert is needed to handle the request
- Leave it unassigned… - List the task as part of the request, but do not assign it to anyone
Once you have created all the tasks required for the specific request type, make sure to press Save!
What's happening in the background?
By assigning a team and responding to a request with an article, Spoke learns to answer just like you do. Spoke works by comparing language and keywords.
Keywords from each team are associated with keywords from each incoming request. Through this matching (and a little machine learning) Spoke learns which pieces of knowledge to use.