Projects
Abzu is a deep-tech startup specialising in artificial intelligence (AI) for medical and scientific research, with a strong emphasis on explainability and transparency. Initially focused on these fields, Abzu aimed to expand its reach beyond healthcare. We began developing an AI-based statistical analysis app as a subscription platform, but after three months of iterations and product research, we realised that a full-scale launch wasn't feasible with our resources. Consequently, leadership decided to pivot towards creating proof-of-concept prototypes to explore opportunities in other sectors and secure further EIC grant funding by demonstrating Abzu’s AI potential across various industries.
Embracing the new strategy, we optimized our workflows to create multiple functional prototypes within a six-month timeframe.
Adapting to our new strategy required rapid development. I conducted a thorough UI/UX audit of our existing platform to assess its salvageability. I identified key issues: our quickly built components using Material Design were not reusable and carried significant technical debt. Additionally, the absence of standard patterns led to time-consuming adjustments for engineers. Although I typically wouldn’t initiate a design system so early in a project without a clear product vision, the need to start from scratch and the expectation of multiple iterations and app versions led me to propose developing a minimal design system. This approach aimed to streamline our processes and accelerate delivery.
I simplified prototypes and implemented a minimal yet robust design system. I focused on essential design tokens and patterns, streamlining component structures and organising our workflow. This empowered us to ship functional demos in 2-week sprints, reduce technical debt and ensure cohesive prototypes with consistent UX and UI. I streamlined collaboration between our design and engineering teams by integrating Figma files with Storybook, bolstering our front-end design system. Additionally, I worked with the front-end team to establish and document naming conventions and product terminology, promoting clarity and consistency across departments.
I created versatile components for seamless UI integration and set best practices for ensuring compatibility with our design library. We slowly updated and applied the new components to existing prototypes and used them to create new ones.
I applied this new design system across all in-development prototypes. The design system ensures cohesive branding and UI, aligning each prototype with Abzu's visual identity for a consistent look and feel. It enhances UX by standardising core design patterns, which improves intuitiveness and interaction. Additionally, the system speeds up development through reusable solutions, boosting efficiency and fostering better collaboration across teams.
In six months, we partnered with leading companies in aviation, pharmaceuticals, and insurance to develop proof-of-concept prototypes utilising real data for predictive and statistical analysis, information retrieval and more. These prototypes include the AI Airport Agents that makes real-time and highly accurate predictions of plane taxi-times and customer service apps that features a reliability score to assess the AI responses’ accuracy and prevent hallucinations. These prototypes were instrumental in securing additional grant funding from the European Innovation Council, enabling us to demonstrate AI's capabilities effectively and refine solutions using our design system.
