Overview:
Graphium is a B2B SaaS platform founded in 2021, designed to help organisations unlock value from large volumes of unstructured research data.
Many enterprises generate vast amounts of research, but much of it becomes siloed, duplicated, or underutilised. Graphium uses AI, machine learning, and graph technology to automatically collate, structure, and visualise this data enabling faster insight discovery and more informed decision-making.
As a co-founder, I played a key role in shaping both the product and the brand from the ground up, with a strong focus on making AI-driven outputs usable and understandable for end users.
My Role:
Co-Founder, Design & Marketing Director (2021–2023)
I led the end-to-end design and product direction, working across product, AI workflows, and go-to-market strategy.
My responsibilities included:
- Defining the product vision and user experience from 0→1
- Designing UX flows, wireframes, and high-fidelity prototypes
- Translating complex AI and data outputs into clear, intuitive interfaces
- Shaping how machine learning insights were surfaced and interacted with
- Creating the brand identity and visual language
- Developing and executing marketing and growth strategies
- Collaborating closely with engineers to align UX with AI capabilities and technical constraints
- Supporting product positioning and go-to-market strategy
This role required balancing product thinking, AI complexity, design execution, and commercial impact.
Tools:
- Adobe XD
- Illustrator
- Photoshop
- InDesign
- Jira
- Confluence
- After Effects
- Miro
- Webflow
- Close
- Teams
- Word
- Keynote
- Semrush
- Vimeo
- Sketch
The Problem:
Large organisations (e.g. consulting firms, government departments, and enterprise teams) face a growing challenge:
- Research is generated faster than it can be used
- Data is stored across disconnected systems and formats
- Valuable insights are buried in unstructured documents
- Teams duplicate work due to lack of visibility
Despite advances in AI, many tools fail to make outputs usable or actionable.
The Challenge:
Designing Graphium involved solving a complex, AI-driven problem:
- Translating abstract data structures and AI outputs into intuitive interfaces
- Making machine learning results explainable and trustworthy
- Designing for multiple user types (analysts, researchers, leadership)
- Balancing flexibility with simplicity in a data-heavy environment
- Creating a product from scratch with no existing patterns
The goal was to bridge the gap between AI capability and human usability.
Approach:
As an early-stage AI product, the focus was on rapid iteration and validation:
- Worked closely with stakeholders and users to understand research workflows
- Identified key pain points in data discovery and knowledge sharing
- Mapped journeys around data ingestion, AI processing, and insight discovery
- Used AI-assisted tools to explore interface patterns and rapidly prototype concepts
- Iterated on prototypes to test usability and refine how AI outputs were presented
- Collaborated with engineers to align UX with model capabilities and constraints
This ensured the product delivered both technical power and user clarity.
Solution:
Graphium enables organisations to:
- Upload and process large volumes of unstructured documents
- Automatically extract and structure key topics using AI
- Visualise relationships through an interactive knowledge graph
- Explore insights and connections across datasets
- Discover expertise and reduce duplicated work
The platform transforms static documents into a dynamic, AI-powered knowledge system.
Impact:
- Designed and launched a 0→1 AI-powered SaaS product
- Enabled faster research discovery and reduced duplication
- Improved accessibility and usability of AI-generated insights
- Increased visibility of internal knowledge across teams
- Established a strong product and brand foundation for growth
My Design Process:
- Early-stage discovery and problem validation
- Mapping AI workflows and user journeys
- Wireframing and rapid prototyping
- Using AI tools to explore concepts, generate variations, and accelerate iteration
- Iterative testing focused on clarity and trust in AI outputs
- Collaboration with engineering to align UX with AI capabilities
- Brand and design system creation
- Ongoing iteration and refinement
Product:
Graphium is an AI-powered SaaS platform that processes unstructured data (e.g. PDFs, documents) to extract insights and generate a knowledge graph—enabling organisations to explore, understand, and act on their data more effectively.
Outcome:
Graphium demonstrates my ability to design AI-powered products from 0→1—bridging complex machine learning systems with intuitive user experiences.
By focusing on clarity, trust, and usability, the platform makes advanced AI capabilities accessible and actionable for real users.
Wireframe:
To see the full user journey and wireframe, please scroll the embedded pdf below or click to enlarge. To see the full high fidelity prototype created using Adobe XD please click on the button at the end of the PDF or by clicking here.
