I’ve been in the software industry long enough to witness several different methods of requirements gathering.
For years, the process looked pretty much the same. Stakeholders would all gather either in person (always better in my opinion) or online to capture user stories, document requirements, and describe the application they wanted built. The task of translating those requirements into working software then fell to the developers.
Sometimes they got it right, while other times they missed the mark completely.
The problem wasn’t a lack of effort or skill. It was that users were trying to describe something that didn’t exist yet, and it could take anywhere from weeks to months (sprints) to figure out everyone was not on the same page.
This gap was partially addressed through wireframing and rapid prototyping tools that allowed stakeholders to visualize applications before development began. However, a wireframe was still just a representation of the finished product. It’s not working software.
I just completed a hands-on workshop with a customer where we used Pega Blueprint to capture the requirements in real time. All the different Stages and Steps did a great job of capturing the current state of the process, but a lot of the nuance of how the process could run was buried in the comments of each Step.
To quickly turn the Blueprint into working software, I needed to get those comments to my developers in the most efficient way possible. I could have manually reviewed each step, copied the comments into a document, and organized everything by hand. Instead, I decided to let AI do the heavy lifting.
Using the Blueprint AI assistant, I asked it to identify every step that contained comments and then generate a formatted table containing the Step name along with the associated notes. To my surprise, the result was exactly what I needed. Within seconds, I had a structured summary of the conversations that would have otherwise taken considerable time to compile manually.
After importing the Blueprint into a live Pega Platform instance, I provided the development team with both the process model and the AI-generated notes. The result was a working prototype in less than a week that reflected not only the customer’s process, but also the context, assumptions, and future-state ideas discussed during the workshop.
For those interested, I’ve included a short video demonstrating how I use Blueprint to capture requirements and then use the AI assistant to extract and organize those insights for the build team.
Cheers