Zenhub AI
When AI meets Agile
Adapt to the market by setting apart from other productivity tools while implementing lightweight and high-valuable features to remain competitive and pioneering in the world of AI for startups.
Summary
After ChatGPT's success, Zenhub embraced GenAI to improve user experience and simplify workflows for software teams. We focused on fostering agile collaboration between developers and managers, reducing meetings, and refining processes with 6 key features.
ROLE
Senior Product Designer
YEAR
2023
TOOLS
Figma, Notion, Miro, Lyssna, Mixpanel and Fullstory
DIRECT CONTRIBUTION
As Senior Designer, I designed two of the three AI features, analyzed user insights from interviews and surveys, and created prototypes, validating them post-launch. I worked closely with product managers, developers, and stakeholders to align the vision for Zenhub's AI features, focusing on lightweight solutions.
Solution
In under 7 months, we launched a functional AI suite in Zenhub, focusing on features that enhance Agile practices. We carefully identified which tasks needed AI to meet user needs, streamlining processes like sprint reviews, suggested labels, and acceptance criteria. This approach made Agile adoption easier, brought immediate value to users, and positioned Zenhub competitively in the GenAI space.
Challenge
Zenhub faced the challenge of reducing churn and staying competitive in a fast-growing market. With the rise of AI, we had to rethink our design approach, exploring how we could add value beyond ChatGPT’s capabilities. As AI quickly gained popularity, we focused on understanding its potential, experimenting with new patterns and UI ideas. Reaching a critical point, we had to decide: implement AI purely for innovation or enhance our existing system. Ultimately, we chose to do both—leveraging AI for relevancy while improving our core algorithms to provide greater value to users.
01.
Boost accuracy, keep trust.
Make better, more informed decisions for your organization knowing your strategies are based on up-to-date, accurate data.
02.
Minimize busywork
Simplify tasks, automate busywork, and free up time for high-value work.
03.
Simplify process
Make AI your on-demand Agile coach in Zenhub, simplifying project management so it doesn't feel like an extra job.
Goals
REPORTS & SUMMARIES
AI Sprint Reviews
Previously, users who wanted to prepare for a Sprint Demo or Sprint Review would need to manually find all the issues closed in the sprint, summarize them and then create a sprint review/demo document, typically outside of our product.
ACCURATE CATEGORIZATION
AI Suggested Labels
Managing labels in project management systems like GitHub is becoming harder due to "label sprawl" and duplicate labels. As teams grow, this creates confusion, inconsistent categorization, and underuse of labels. As a result, project managers have to manually apply labels, wasting time and making it harder to track progress effectively.
SIMPLIFY PLANNING
Generate AI Acceptance Criteria
Teams often find it difficult to create clear and consistent acceptance criteria for tasks or user stories. This causes confusion, misalignment between developers and stakeholders, and reduces project efficiency. Writing acceptance criteria manually is time-consuming and can lead to incomplete or unclear requirements.
THE RIGHT AI SOLUTIONS FOR REAL PROBLEMS
Why AI and Agile in the first place?
When building AI features, we knew it was important to stay focused on delivering real value, and our goals were pretty ambitious. The key? Listening closely to our users—loud and clear.
Agile methods helped us tackle this challenge and stay flexible, making sure we were always moving in the right direction.
The process
THE STORY BEGIN
Harder, Better, Faster, Stronger
Building AI features from scratch during the ChatGPT boom was quite a challenge...
01. Research & Conceptualization
We kicked things off with quick, lite sprints and ideation sessions, bringing developers and designers together. Then, we sent out surveys to understand what our users expected from AI. This was crucial in validating our vision and making sure we were on the right path.
02. Product Vision
Not everything was easy or doable, but as a team, we built a clear product vision that would guide us forward. Using data, we identified where AI could make a real impact while staying aligned with our OKRs and KPIs. Workshops with stakeholders helped turn this vision into achievable milestones.
03. Benchmarking & AI patterns
The AI boom was happening everywhere, and we knew we had to keep up. I dove into learning new patterns, taking courses, and studying the competition. As a startup, we didn’t have the same resources, but we aimed to stand out by doing things our way.
04. Prototyping & Quick iterations
Instead of starting with sketches, we jumped straight into mock-ups and integrated them into our design system. Thankfully, much of what we already had was reusable. Along the way, I picked up some neat prototyping and motion design tricks to bring our vision to life for stakeholders.
05. AI Opt-in Testing
Betting on AI was a big goal for the company. There were wins and losses, but we managed to get users to opt in and test AI features. This gave us the feedback we needed to make weekly improvements and build a better experience.
06. Development Hand-off & Trade-offs
Building GenAI was a whole new challenge, everyone had ideas, but execution was tricky. I worked closely with developers to find a balance between maintaining the core idea and not sacrificing performance. It was all about keeping quality intact while moving fast.
The impact in numbers
3
AI features rolled out in less than 6 months
32%
WAU Increase in 3 months
7.8%
Conversion Rate increase
Trust and adoption
take time
Even with quick implementation, users remained skeptical of AI. Building trust through consistent value and clear communication is key to increasing adoption.
Collaboration across teams is key
Successful delivery relied on close teamwork between design, development, and product teams, making trade-offs without compromising quality.
Balance between Innovation and simplicity
Integrating AI added value, but users still preferred simple core tasks. We learned that not all processes need AI, so it's important to prioritize user needs over trends.