At Hey Nowinnovation is lived in the products we develop for our customers, and also within the organization. As part of our commitment to improve our internal processes, we developed a bot designed to assist product owners (PO) in the creation of user stories (US) and product backlog items (PBI), within the framework of best practices of agile methodologies.
The context
The role of the product owner is essential in agile teams. Their ability to structure ideas and requirements effectively can make the difference in the success of a project. With this challenge in mind, we applied our knowledge and experience in AI to create a tool that would simplify and streamline backlog management.
The process
Identification of the internal need
We noticed how PBI creation and backlog management can consume valuable time in our projects, so we set out to optimize this process.
Customized development
Using technologies such as Claude and ChatGPT, we design a flow in which the bot automatically generates features, US and PBI from documents (such as PDF, images of user flows) or results of techniques such as brainstorming.
Tests and adjustments
We tested the bot in real situations and adjusted its functionalities based on feedback from colleagues to ensure that it would be useful in our agile environment.
Designing the bot
AI models
We use Claude and ChatGPT because of their ability to interpret natural language and adapt to the specific needs of our workflow. In particular, these models are useful in agile methodologies contexts, as they allow a quick generation of detailed and structured user stories, with clear and concise acceptance criteria.
Bot training
Using historical data from previous projects, user stories, PBI and acceptance criteria, we train the model to understand the language and specificities of our tasks, to ensure consistency and quality of results.
Iteration and feedback
During internal testing, we received continuous feedback from product owners and developers. This iteration process was key to fine-tune the bot's responses and improve its accuracy.
Example of response to ambiguous functionalities
From the feedback, we adjust the bot to ask clarification questions when it detects ambiguities in the inputs. This improves the clarity and specificity of the BIPs.
Results
The bot has facilitated backlog management and allowed teams to focus on higher value tasks.
Reduction of product owner management times
After implementing the bot, we were able to reduce the time spent by POs creating and refining PBIs and US by 40%. This means that activities that used to take several hours can now be completed in a fraction of the time, allowing teams to focus on more strategic and creative tasks.
25% increase in productivity
Through automation, we have seen a 25% increase in productivity, which is reflected in fewer refinement meetings.
Improved quality of deliverables
Thanks to automation, PBIs also reduced errors and misunderstandings by 20%, resulting in fewer review cycles.
Reflection
These types of initiatives demonstrate that innovation does not always have to be external; it can and should start within the organization. We are excited to continue exploring new ways to improve our work through technology.
By:
Iván Coronel, Product Designer
Ivan has more than 8 years of experience in digital product design. Specialized in interface design and agile methodologies, he has incorporated artificial intelligence to optimize processes. In his role, he turns innovative ideas into effective solutions.
