💡 How to automate customer service with Make, Jira and ChatGPT

Hey Makers :wave:


Staff shortages are a real drag.

Whether your company is short on human resources, or there’s some seasonal flu keeping employees home, it can be tough to get by without enough people. And, customer service is one department that doesn’t get to take a day off.


:brain: IT consultancy DEMICON GmbH solved this very problem for its client, a car parts supplier, who was suffering from both a general employee shortage and sick-leave absences.

They turned to Make to integrate (what else!) ChatGPT with the company’s Jira Service Management platform. By leveraging the power of AI, they were able to collect key data from customers and instantly route it to the right agents.

:rocket: The results, as you might imagine, have been impressive:

  • 20-30% reduction in service agents’ workload
  • Improvements in data classification
  • Increase in service quality and productivity
  • No need for new staff or increased spending
  • More accurate reporting.

:point_down: Keep reading to discover how AI technology can help revolutionize customer service.



What problem were you trying to solve with your automation?

The first level of the service desk was understaffed due to many agents’ sick leave. We automated the first contact for the customers.



Why did the problem exist?

Not enough employees, many absences due to sick leave, many service requests.



How did you solve the problem? What does your solution look like?

Integrating ChatGpt with Jira Service Management. The first contact from the creation of the issue throughout the 3-5 questions to the help-seeking persons is done with AI ChatGpt provides.

What the scenario does:

  • A webhook from Jira triggers the scenario and an issue gets created.
  • Based on the summary, the OpenAI endpoint is asked for guidance and the answer is delivered to Jira as a comment to the customer.
  • If the customers replies, the scenario is triggered again.

a peek behind the automation curtains



What did your solution achieve?

  • Approximately 20-30% reduction in manual workload
  • More satisfied customers - metrics for “Time to 1st response” show better numbers
  • Higher solution rate with better-qualified tickets.

Christian Wiemer, Demicon’s Solution Lead and the man behind the automation



Helpful Resources

:make: OpenAI (DALL-E & ChatGPT) on Make
🟣 Getting started with OpenAI GPT-3 and Make
:make: How to Automate your Workforce with AI and Automation

2 Likes