Second Level Automation (L2)
Having recognized whether a request is an incident or service request, some customers may want further sub-classifications, i.e. automation of a second level classification.
Aportio is developing and extending standard classifiers to deal with common IT request types. These are being progressively created and deployed and provide opportunities for service providers to further refine how they action common requests from customers.
Common second level classifiers
Classifier | Nature of question | Potential response |
---|---|---|
*User configuration | I want to add, change or remove a user or their system access rights | present a form to capture information |
*Distribution lists | I want to add or change email distribution lists and memberships | present a list of clarifying questions |
*Email diagnostic | I have a problem with accessing email or mailboxes | present a list of knowledge article links |
Password reset | I have a login or password related issue that requires a password reset | Ask for user email address and mobile to authenticate user |
Network diagnostic | I have a network access or network performance problem | clarifying questions on extent of impact, and when issue started |
Email changes | I want to add or change user access rights to email or mailboxes | present a pick list of common options |
Mail security management | I want to black or white list email addresses or release or quarantine emails | trigger a 3rd party application to select/release/action |
Application rights | I want to add new access rights to an application | present a list of clarifying questions |
Procurement | I want to buy a new device of some sort | assign to a product procurement team to action |
*Available in standard Level 2 configurations. Ask us for a quote for other L2 classifiers.
Combining multiple classifications into powerful solutions
A hybrid approach example uses:
- conditional constants (e.g. a persons job title)
- regular expressions (e.g. a specific service name)
- heuristics (related word patterns)
- NLP sentiment analysis (emotion). The email is tagged with the appropriate classification using:
The InboxAgent solution combines Artificial Intelligence/Machine Learning with one or more trained other techniques to classify and action emails.