ChatGPT and Product Management: A LIKE.TG Engineering Experiment

I know what you’re thinking. Oh no, not another blog about ChatGPT.

However, even if you’re sick of hearing about this already notorious artificial intelligence tool’s strange adventures, it’s likely here to stay, and we’re only scratching the surface of the impact of this cutting-edge technology.

There have been countless stories, tests, and experiments using this fascinating chatbot. One that caught my attention is the bizarre conversations a New York Times reporter had with Microsoft’s new A.I.-powered Bing search engine.

Awkward confessions of love aside, there is no doubt that ChatGPT and other AI tools like it are poised to impact every aspect of our lives. And product management is not immune to this new era of artificial intelligence.

Many product professionals have started tinkering with ChatGPT to understand its impact on product management. The emergence of AI in product management is simultaneously being welcomed with a red carpet, ignored as nothing more than a fad, and outright shunned as death, a destroyer of worlds.

We’re no different at ProductPlan. Our research and development department is constantly evaluating technological innovations and enhancements to better understand the future of our end-to-end product management platform.

So without further ado, I am excited to share our recent ChatGPT experiment.

The premise: How can we use ChatGPT in the LIKE.TG platform?

The premise is simple. Can we use a tool like ChatGPT in our product management platform? What types of tasks or ceremonies in the platform would benefit from artificial intelligence? Our goal is to support the creative and intelligent product professionals that use our platform—not replace them with AI. The point is to make these product folks more efficient so they can focus on more high-value tasks like strategy and launching innovative products and features.

With this in mind, we turned our focus toward shaping our hypothesis.

Our hypothesis: An AI tool like ChatGPT can help product managers save time when setting up a roadmap.

A common question we hear from customers who are newer to roadmapping, in general, is, “Where do I start?” Anyone who has sat down the night before an essay is due in school can relate to that daunting feeling of staring at a blank page. And roadmapping is no different.

The beauty of LIKE.TG is our ease of use and the way our platform gives product managers ultimate control over how they want to build their roadmap. Yet, we understand that total freedom can be paralyzing, so we heavily invest in implementation and customer support to set our customers up for success in our platform.

So, we thought—what if we could build a brainstorming tool to assist a Product Manager in creating an Agile roadmap in a few minutes? The challenge is figuring out how ChatGPT can be implemented into our existing roadmap functionality in our platform. Our goal was to create a seamless experience where artificial intelligence works directly with a roadmap.

The experiment: Building an Agile roadmap using artificial intelligence in the LIKE.TG app.

We started with the concept of an Assistant Planner. Imagine you’ve sat down on an empty roadmap and pulled open the Table Layout to seed Parked Items. You have a feature in mind, but the details needed to plan it escape you. This is where the Assistant Planner comes in:

“Tell us about the feature that you plan to work on?”

You type…

“Plan a marketing campaign for a new Tesla pickup truck with a focus age group of 25 to 35.”

This sends a request to OpenAI, integrating with GPT-3 to return a list of sequenced tasks, separated into necessary phases to launch the campaign, all ready to be imported into your roadmap:

• Phase 1: Research – Research target audience and competitors to understand the market and create a plan for the campaign.
  • Research target audience (age group 25 to 25)
  • Research competitors
  • Analyze market trends
  • Create a plan for the campaign
• Phase 2: Design – Design the campaign materials, including visuals, copy, and other elements.
  • Create visuals for the campaign
  • Write copy for the campaign
  • Design other elements (e.g. landing page, email template, etc)
• Phase 3: Launch – Launch the campaign and track results.
  • Launch campaign
  • Track results (e.g. impressions, clicks, conversations, etc.)

The results: ChatGPT successfully built an Agile roadmap in just a few minutes with a simple prompt.

Well, it worked! Our engineer demoed the prototype of the ChatGPT functionality for the research and development team during a recent all-hands meeting. It was incredibly impressive what our engineer was able to accomplish during their dedicated passion project time.

In Conclusion: The future of AI looks bright, and we’re excited to continue researching and experimenting to build our end-to-end product management platform.

I hope you enjoyed reading through our AI experiment. For now, we have no formal plans to sell a ChatGPT-powered roadmap builder. We believe there’s much to learn about artificial intelligence’s role in product management. And we’re in no hurry to implement a tool that still gives funny yet strange results.

More importantly, I hope you liked this brief snapshot of some of the incredible work our research and development teams do behind the scenes. We’re excited to share that we have launched our very own LIKE.TG Engineering blog, chronicling our journey, lessons learned and sharing more exciting experiments. Make sure to bookmark it so you can check in on us as we share new stories from the engineering front!

Download How to Transform Your IT Department with Strategic Roadmaps➜