Summary
Role: Lead UX & Product Designer (Consultant) | 0→1, End-to-End Design
I led the concept and product design for UBIX.ai, an AI-powered Big Data Analytics platform that combines Generative AI and Reinforcement Learning to give business analysts the power of a data scientist.
My work helped launch the V1 platform, secure the next funding round, and define AI UX patterns that are widely used today.
The Challenge
UBIX.ai aimed to democratize big data analytics with a conversational AI interface. I joined as one of two designers to:
- Create clarity in a highly complex and ambiguous problem space
- Design the first end-to-end experience for V1
- Innovate new UX patterns for prompt-based AI applications
- Support investor pitches and fundraising
1. Understand Users and Workflows
- Partnered with the founders, data scientists, and engineers
- Created personas and scenarios based on real data science workflows
- Predicting flight delays
- Oil field predictive maintenance
- Cancer genome analysis for treatment planning
- Predicting global shipping rates
Personas
Scenario: Oil field predictive maintenance
"Imagine it was human". That principle guided us in creating the scenarios that kept the team aligned and informed the design throughout the process. We analyzed conversations between data scientists and their stakeholders and identified common tasks they performed, their workflows, and how they arrived at answering questions in ways that were actionable for the business.
2. Explore & Validate Concepts
- Whiteboard explorations to translate scenarios into potential interactions
- Wireframes and flows to test ideas and collect team feedback
- Focused on core challenges:
- Guiding first-time users with prompt suggestions
- Navigating branching conversation histories
- Designing specialized visualizations for complex data (e.g., Onco Plots)
Whiteboard Sessions
Wireframes
3. Prototype for Investors & Launch
- Built clickable prototypes for investor presentations and demos
- Created screen flows that communicated the user journey clearly
- Delivered final designs that were implemented in the V1 SaaS platform
Screen Flows
Key Design Highlights
During his project, I came up with a number of design patterns that have since become universal AI design patterns.
Prompt Guidance
The empty prompt box that is typical for conversational AI application makes it daunting for users to get started. Examples of suggested questions and conversations give users a sense of the kind of prompts they can use to get started.
Conversation History Map
AI conversations can generate large amounts of content quickly. Working with the team, we observed that in data analysis workflows, users would often go down a path, find out it didn't lead to the desired results, and then go back to an earllier point in the flow and pursue a different branch of the analysis. I designed a tree based conversation history that reflected that. The tree can be navigated via pan and zoom, like a map, allowing users to branch, revisit, and explore alternate analytical paths.
Advanced Data Visualization
Onco Plot depicting the genetic landscape of bladder cancer.
Investor Deck & Prototype
The designs were used for the schematic in the investor deck and contributed to raising the next round of funding.
Outcomes & Impact
- UBIX launched its V1 SaaS platform
- Secured the next round of funding
- I created reusable AI UX patterns that remain relevant years later
- Gained deep expertise in designing for expert workflows and AI-first products
Takeaway
Designing UBIX.ai was a true 0→1 experience, transforming a sea of technical possibilities into a clear, actionable, and fundable product experience. It reinforced my belief that grounding innovation in real user workflows and designing with first principles in mind creates lasting value.
