ChatGPT and DevOps: Practical Uses for Disruptive AI

Reading Time: 3 minutes

If DevOps’ mission is to increase the speed and quality of application and service delivery, is it possible for DevOps practitioners to get the benefit now from a new, disruptive artificial intelligence—ChatGPT? The answer is a qualified “yes.” In this blog, you’ll discover how ChatGPT can be used to automate technical documentation creation and learn how other technologies can support DevOps in complex enterprise environments—the way TIBCO customer Euler Hermes has done.

What is ChatGPT?

For those who’ve not been following the tsunami of articles and social media posts about ChatGPT, here’s a very brief overview. ChatGPT is a Large Language Learning Model (LLM) plus chatbot being tested by OpenAI, an organization focused on developing Artificial General Intelligence (AGI) technologies, funded by Microsoft and other sources. One use case discussed for ChatGPT seemed particularly relevant for DevOps—that of automating the creation of technical documentation from technical specifications and artifacts. Since technical documentation can help speed the adoption of new applications, it’s key to a robust DevOps practice. 

ChatGPT Put to the Test

Dave Winstone, Solutions Engineer at Cloud Software Group, and I decided to see how ChatGPT might work to speed up technical documentation creation. We applied it to a novel TIBCO BusinessWorks use case. In his personal blog, Dave stated he wanted to achieve a Zero Downtime upgrade when a TIBCO BusinessWorks application/microservice is deployed in a Kubernetes environment. Dave said, “When Lori approached me about a potential ChatGPT use case that could help DevOps, I was really quite intrigued. I’d explored ChatGPT for fun family projects, but not to see whether it could be a time-saver for real-world use cases.” 

To test the idea, Dave and I separately worked with ChatGPT. Here are the steps we followed:

  1. Open ChatGPT, and type in a prompt that requests the creation of technical documentation.
  2. Scrape the blog contents into the ChatGPT interface beneath the typed prompt. (Note that images were not able to be consumed, and code snippets needed to be manually copied.)
  3. Submit the prompt and contents.
  4. Review the response, resubmit, and prompt for improvements.
  5. Copy the result out to a template for technical documentation, review, edit, and manually add screenshots.
  6. Review, run QA, and publish.

Time saved: An estimated 10-20 minutes, depending on the editing support needed.

Although the amount of time saved in this early ChatGPT experiment is modest, it indicates a seachange in documentation creation is coming, and the time to experiment is now. When multiplied across an enterprise—where DevOps is speeding delivery on a myriad of applications and updates—augmented documentation creation will be a DevOps force multiplier. Despite the fact that ChatGPT is still in its early days, we can expect its usefulness to increase rapidly as more users engage and thus help tune the model. But DevOps practitioners need not wait for novel technology to improve DevOps outcomes. There are proven technologies with DevOps benefits, such as the subject of Dave’s blog, TIBCO BusinessWorks (available within the TIBCO iPass connectivity platform). 

"Although the amount of time saved in this early ChatGPT experiment is modest, it indicates a seachange in documentation creation is coming, and the time to experiment is now." Click To Tweet

Accelerate Your DevOps with TIBCO

Technologies like TIBCO BusinessWorks support cloud-native continuous delivery models, key to modern DevOps. Using continuous delivery, IT can reduce development cycles to short sprints, incorporating small, incremental changes into an application frequently, even iterating several times a day. 

Guillaume Compastie, IT project manager at Euler Hermes, describes the acceleration possible with TIBCO: “Some years ago, good project completion was one or two years. Today, people expect it in six months, and probably in one or two years, they will expect projects within three or four months, maximum. And for that, you need the right architecture. You need an ESB and the ability for flexible management of transactions.”

The combination of a DevOps culture, microservices, and enterprise platform as a service (PaaS) enables the responsiveness needed to keep pace with change. And even if it’s a little bit early days, in keeping with the DevOps focus on acceleration, the experimental automation of technical documentation creation through ChatGPT is worth consideration. Let us know how your ChatGPT experiments go and how TIBCO can help your DevOps acceleration.

Learn more about TIBCO Integration Capabilities and DevOps.