Photo by Levart_Photographer on Unsplash
We have been following the buzz around ChatGPT along with everyone else, and had some internal discussions about how we might use OpenAI with Request Tracker. If you have somehow missed all of the news around OpenAI, we decided to ask ChatGPT to introduce the topic for us, and it came up with this:
Organizations are looking for ways to improve their customer support and ticket management processes as the world becomes more connected and customer expectations continue to rise. One way to do this is by leveraging the power of AI and machine learning. OpenAI, the company behind ChatGPT, provides AI technology, including language generation and natural language processing algorithms that can make Request Tracker (RT) more powerful.
OK, that's pretty good for an AI, but how would you actually use it? As a start, we created a simple action in RT to call OpenAI and perform sentiment analysis on the contents of example support emails. Based on the result, we thought we might better understand how a customer might be feeling.
What is Sentiment Analysis?
Sentiment analysis is the process of analyzing text to determine the emotional tone or attitude of the writer. It can be used to identify whether a customer is expressing positive, negative, or neutral sentiment, and can be a valuable tool for improving customer support and engagement. Sentiment analysis can also help organizations identify trends and patterns in customer feedback, which can be used to inform product development, marketing campaigns, and other strategic initiatives.
Edit: 6.14.23. It’s worth noting that in the ever-changing AI ecosphere, any number of concerns about the safety and uses of AI technology have arisen from more than a handful of communities. Today, the European Parliament adopted its negotiating position on what it calls the Artificial Intelligence Act.
Generally speaking, the new rules would ban AI systems for social scoring, biometric categorization, and emotion recognition. Negotiations with the Council on the final form of the law will begin later today.
Integrating OpenAI with RT
To integrate OpenAI with RT, we can added a custom scrip to run when a customer replies on a support ticket. We use the Perl module OpenAI::API to call the OpenAPI service and get a sentiment value back: positive, neutral, or negative. Then we store it in a custom field for easy reference. Here's a video showing our demo in action, testing with a few different types of replies.
Calling OpenAI from RT
Doing more with OpenAI
Once you have the sentiment value, you can think about other ways you might use it. The simplest is reporting, allowing you to see how customers feel at the end of their support interaction. But you could also take action, like routing "negative" tickets to more experienced staff or following up later with customers who had their question answered, but remained with a "negative" sentiment.
And sentiment analysis is just one of the many things you can do with OpenAI. You could ask it to summarize long support questions, suggest possible solutions, or maybe provide a summary of previous support interactions with the same customer.
Interested in using OpenAI with your RT?
This was a simple demo we did as we saw all of the news about ChatGPT, but there are definitely some real-world use cases for this type of technology. If you'd like to get more information on this or want to discuss how leveraging OpenAI with RT might be beneficial to your organization, send us email at sales@bestpractical.com and we'd love to "Chat".