Answering queries from investment analysts about companies using GPT-4 and Portrait Analytics technology

Team & Role

As the first product designer, I collaborated with the two startup founders and four engineers to prepare the second iteration of the product.

What I Did

Working alongside the founders, I assisted in revamping the platform and developing a design system to enable front-end engineers to construct the new demo. This is in line with the company's goal to grow and attract investor attention during their initial round of funding.

What I Delivered

Competitive analysis, task analysis, user flows, wireframes, engineering coordination, usability tests, voice and tone analysis, design system, and clickable prototypes.

Project Outcome

I facilitated the process of obtaining seed funding by assisting the founders of Portrait Analytics in finalizing the product definition through a series of user experience (UX) activities. I collaborated with engineers to enhance the look and functionality of the new demo, gathered early feedback from users, established a consistent design aesthetic through voice and tone exercises, and built a design system from scratch – all at light speed.


By analyzing the competition, we identified a need for a financial document management tool with a smart search engine that prioritizes relevance over chronology, and easy navigation between documents. Existing tools often have poor search capabilities and a cluttered interface, making it difficult for analysts to efficiently research companies and identify new investment opportunities. 

Analysts lack a user-friendly dashboard that streamlines their research process and provides them with the confidence they need to make informed decisions in record time.


To determine which design elements and features to incorporate from our competitors and which to avoid, I reviewed a series of think-aloud exercises with the founders. We examined three tools: BamSEC, Sentieo, and Alphasense. 

Based on this analysis, we established the desired flow of the platform. The process should start with analysts selecting the company they want to research, selecting the financial documents they will use as a basis for their research, and defining a time frame. Then, they can access a filtered list of documents and use our search engine to ask questions.

Voice & Tone

"...there is a confidence in being casual –but not unprofessional– that we want to internalize." Portrait Analytics Founder

To establish the identity of Portrait Analytics and ensure that our aesthetic decisions align with our desired image, I conducted a series of voice and tone exercises with the company's founders. We decided to present ourselves as professional and dependable, but without coming across as overly formal or intimidating. Our aim is to be like a human analyst - enthusiastic about our research, focused on providing valuable insights, and dedicated to delivering exceptional results.

Responsive Layout

With remote work on the rise, I focused on ensuring seamless usability of our dashboard when transitioning from a larger screen to a smaller device such as a MacBook.

Through implementing a flexible six-column grid and strategically allowing for some elements to adapt in size while preserving key functionality, I was able to maintain the balance and user-friendly design of the dashboard.

Engineering Coordination

"To hit our deadline for this round, we need to streamline the work for GPT-3…" Backend Engineer

The design process wasn't without its challenges. Collaborating with back-end engineers affected both the look and performance of the design. For example, when designing the crucial "Ask a question" search bar, the back-end engineers requested that I simplify the task for GPT4 by including a set of predetermined questions that aligned better with their document parsing methods for GPT4’s consumption.

Other work

Want to create user-centered solutions? Drop me an email.