Risk/Impact matrix for automation
In a support team, one of the critical aspects of automation efforts is defining what you automate and how. In Lang.ai, we are working with companies using Zendesk to help them figure that out, and we have designed the following matrix as a strategic framework.
A risk matrix is a matrix that is used during risk assessment to define the level of risk by considering the category of probability or likelihood against the category of consequence severity. It is a simple mechanism to increase the visibility of risks and assist management decision making.
We have applied it to support to help teams set up a strategy for automation and we are sharing our method so that any support team can apply it.
Low risk/low impact:
These are ticket topics that are not very frequent and also aren’t very risky for the company. This bucket typically includes questions about product usage, that you could potentially redirect to a well-designed help center with automated responses. These are the best topics to start playing with when you want to automate.
By automating them, customers will solve their problems faster and have a better customer experience than if they have to wait for an agent to respond.
As these topics are not frequent and there may be many, it may be time-consuming to build a help center article for each of them, but you still should apply SLAs. You can also assign low priority.
For instance, a key area of your product may be a higher priority than a less important question about something tangential.
Automation recommended: Set a low priority
Automation recommended: Apply baseline SLAs
Low risk/High impact:
These are also ticket topics that have a low risk for the company (questions about product usage or similar) but in this case, they are very frequent, so the impact is really high. It’s critical to have help center articles for these topics and redirect the feedback to the product teams as there may be a bigger UX problem in the product. Product teams should be closely looking at these issues.
Automation recommended: help center redirection/automated response.
Automation recommended: triage to product teams
High risk/Low impact:
These are ticket topics that are not very frequent but have a huge impact. An example is for a grocery delivery company, finding a bug in your food, for a SaaS company someone not being able to use a core functionality of the product (e.g., you cannot send an email in Gmail). They are of low impact because it may be only a limited set of users or tickets. It is essential to solve these tickets ASAP and carefully as these frustrated users don’t want to be talking to a bot or redirected somewhere else
Automation recommended: set a high priority
Automation recommended: triage to expert teams
High risk/High impact:
These are ticket topics that are very frequent and have a huge impact. An example is a significant disruption or bug that impedes the use of your product. It is very important to solve these tickets ASAP and solve them correctly. You want your best agents on this responding as quickly as possible to provide the best customer experience.
Automation recommended: set an urgent priority
Automation recommended: triage to your best agents
To find which topics belong to each square in the matrix, Lang.ai can automatically find concepts that are relevant in your Zendesk tickets and replicate your existing tags+ suggest new tags.
This framework was built together with Charlene as part of Lang.ai consulting services offered for our clients. Contact us if you want to learn more about how we can help you improve your customer support operations.
Check the other articles in our Building Lang.ai publication. We write about Machine Learning, Software Development, and our Company Culture.