Pattern: Smart Response to Email Inquiries (Managing FAQs)

Thomas_937381 Posts: 196
edited July 2020 in Show and Tell

This four-step workflow starts via an email trigger. An internal or external customer can email with a question (in the body of the message), and the workflow will send an email response with the answer. If no answer is found, the requester is notified, and prompted to reply with a different question.

This could be especially useful for providing an initial response to FAQs within a support or service organization, and could be integrated with a ticketing system, should yours have an available API.

The workflow primarily leverages Tables: Find similar text. The table contains three columns, Question, Keywords, and Answer. In some cases with a pattern like this, just Question and Answer columns may suffice. Having the Keywords column allows one to include terms used in different variations of the question.

To use this workflow in your team, you'll need to:

  1. Create an email trigger, question @ {yourteamname}
  2. Go to the list of associated Data Tables on the Settings screen, and take note of the default questions (there are four).

If you'd like to test this workflow, try an approximation of one of the below inquiries in an email to question @ {yourteamname}

  1. How do I sequence actions in my workflow, so that they start in a specific order?
  2. What type of action do I need to use if I want to gather data from people?
  3. What if I want separate paths in my workflow to start only if conditions are met?
  4. How can I create a loop in my workflow to return to an earlier step?

Building a workflow like this is simple. The more difficult portion of the exercise comes in the form of studying data from your ticketing system or email inbox, and categorizing repeating inquiries. Once you've done that, you can study the patterns to determine how your customers are describing their questions, and build a model and set of keywords for each. Focus first on those that are most common questions.

Tables: Find similar text gives a Top Match Similarity score, which you should use to continually optimize your questions and answers.