This startup is still in stealth mode, so I cannot mention their name. I also changed some of the copy in the product to conceal their identity the specifics of the product. This company is focused on building fast, accurate, and compliant enterprise AI. Their product enables data science teams to seamlessly build, deploy, and operate machine learning solutions at scale.
I worked on this project as a freelance UI designer. My role was to translate the existing user experience into a smooth and intuitive interaction to share with investors and potential customers.
My Design Process
To ensure that I designed the best possible solution, in the best possible way, I followed the process below because it allowed me to build empathy for the target users and iterate quickly:
Step 1: Empathize
The first step in my process was to understand the problem and the users. Our target audience was data science leaders and data scientists. I needed to understand the way they do things and why, their physical and emotional needs, how they think about the world, and what is meaningful to them. All of this gave me clues about what they think and feel and helped me learn about what they need. Due to resource and time constraints, I was unable to conduct a research study. So I used feedback and data that the founders had already collected from end-users and then turned to Google for the rest. Although this wasn't ideal, you have to be scrappy when working for a startup.
Step 2: Analyze
Now that I had a better understanding of AI and gained invaluable empathy for data scientists and their leaders, I needed to synthesize my findings into powerful insights. I organized, interpreted, and made sense of the information that I gathered to create a problem statement.
Next, I developed personas to summarize the data I collected about our users, what they need, what their goals are, and what they expect. The purpose of these personas was to create reliable and realistic representations of our target audience that I could use throughout the rest of the design process.
Since I was so new to the AI space, I decided to take it a step further and create empathy maps of our users. This technique helped me to gain a deeper understanding and draw out unexpected insights into our user's needs. I wanted my designs to be grounded in a deep understanding of the people for whom I was designing for, so developing this sort of empathy was crucial.
Step 3: Ideate
Now it was time to transition from identifying problems to creating solutions for our users. I started with sketches which allowed me to explore different design approaches in a short amount of time.
Once I felt like I had found the best solution, I created wireframes to better communicate my design. I shared these wireframes with the team for feedback. Once all stakeholders agreed on one design, I was able to start creating high-fidelity mockups.
Step 4: Design
Now that I had finalized the layout and flow of the interface, the final step was to create high fidelity mockups. The startup had previously hired a freelance brand designer to create presentation assets. I decided to use the green in those assets in my design to maintain brand consistency. I chose to use that color for the logo, buttons, to indicate current navigation location, and as the primary color used in data visualization.
The high-fidelity mockups that I designed were used to sell the product vision to VC firms to raise capital. The mockups were also used during meetings with existing investors to update them on the platform's progression.
During this time, the founders were starting to speak to potential beta users. It was vital for them to recruit beta testers to get product validation, collect feedback, and begin proving traction. My mockups were used during these pitches to communicate the problem and what they built to solve for it.
I learned a lot while freelancing for this AI Startup. A big challenge with working for an early stage startup is the limited time and resources. There is a lot of pressure to start sketching out solutions right away. However, it was crucial to contextualize the problem and understand the target audience first. Without building a rationale behind the problem, the reasoning behind my design decisions would have ended up being part of a non-existing framework. So I focused on making sense of my research in the beginning. Doing this allowed me to back up my design decisions with customer insights and data.
Finally, as the only designer, I had no other designers with whom I could bounce ideas. The lack of feedback from another designer forced me to be more critical about the design decisions that I made. I learned to question myself more so that there was always a reason behind everything that I designed. This enabled me to consider how each of my designs decisions addressed the target audience's problems.