Consigliere AI
Creative Director
Brief Overview
Consigliere was an AI/ML tool which utilized machine learning to work alongside influencers, offering unprecedented solutions to improve viewer retention rates.
The startup included a small team of 6 people including myself. Conducted at Northwestern University "The Garage"
Richard "Richie" Wu (Founder), myself (Creative Design Lead), Josh (Software Engineer), and Autumn, Sparsh, Shwet (Back-end)
Initial Phases
Before beginning anything, Richie and I had a Zoom meeting to discuss his vision for the product, goals, and how the program worked. Knowing this I began sketching everything out on a moodboard. Establishing typography, color palettes, inspiration, etc. At the time of joining, they had the back-end setup but the LLM still needed tuning
We had only 3 months of being able to work in-person until the program ended at The Garage. Knowing this I got to work right away.
"I met so many creators and we had many talks about how to best position our content.
If we had something that was smart enough to see small changes in visual components and retention, we would be getting into this game with a lot more intelligence,"
Founder, Richard Wu
The Problem
We wanted to solve the issue of smaller-sized content creators who had difficulty maintaining viewer retention rate consistently or finding what drove it in the first place.
Consigliere would be their advisor within the creator space.
Features
Key features of the program was the script optimizer, thumbnail analysis, and calendar planner.
While it wasn't perfect, it was functional. Which was enough for refinement and revision.
Defining Stakeholder Value
Ideation Wireframe Mockups
Determining Sections/Layout
I put those sketches and concepts into a digital format within Figma while also drafting alternate concept iterations.
Ultimately refining the initial designs later after conducting tests and insights with other users.
User Journey
I began drafting a user journey map, determining individual steps a user might take, varying scenarios for each one. This part of the research is what helps to understand how the user may feel, pain points, and drive solutions.
It's also a great way to visualize and break down the process.
Heuristic Evaluation
Using the first rendition that had already been published before I joined the team, I began going through and marking key areas of the site which need refinement.
Finding a majority of the problems lied in navigation and keeping users informed of the system status.
Usability Testing
Using the data I gathered and after familiarizing the current iteration, I began running usability tests to see how users were to interact with the product.
This would verify two things: One, that the user journey map is (somewhat) accurate and two, see areas which need to be prioritized most.
After running tests and evaluations, we found that a majority of participants were confused within the script optimizer, which failed to show which parts of their script was changed or took too long to analyze.
Refinements
Hi-Fidelity
After compiling all the data from our participants, researching common patterns and faults within the site, I designed a more streamlined version of my initial mockups in addition to onboarding.
However, there was one hurdle I had to overcome.
Because our team was busy working on the back-end and refining the model to ensure everything ran smoothly, I had been asked to code the designs of the UI in myself.
I was quite familiar with HTML, CSS, and Javascript but learning how Tailwind-CSS in a short span of time was something new I had to overcome.
With the best of my abilities and utilizing ChatGPT-4o mini, I was able to incorporate some elements of my design to the front.
Screenshots of Updated Design
Additionally, in order to ensure Consigliere worked, we created an entirely new Youtube channel "Siilken" to see how it would do. Using Richie's voice acting and utilizing my skills in motion graphics + editing, we uploaded videos which were optimized using Consigliere.
Within the first 5 days, the channel had gained over 1.1k new subscribers and 42k views.
Closing Remarks
Overall, working at Consigliere was an eye-opening opportunity. It taught me the hard truth when it comes to startups and the process when dealing with real clients.
A majority of startups often fail. And that's okay because there is no such thing as success without failure.
Initially I had joined the team trying to follow the same process I would for a school class. But after a few weeks of talking with advisors within Northwestern as well as stakeholders, I realized that what I had been taught in class would not work. I couldn't simply follow a structured step-by-step playbook and had to adapt on the fly. Learning ways to conduct tests and evaluations was only one part of being a UX/UI designer
Additionally, having skills in other areas such as motion graphics, photoshop, and graphic design, allowed me to be extremely flexible within my role as Creative Director, which I hope to get better at and learn more with future work.
And while I had to step away to focus on academics, I am extremely grateful to Richie and the rest of the team who helped bring the vision to life and allowed me to be apart of the team. Because of this opportunity provided by Northwestern University, I not only learned more about the business world but also what it takes to be a successful designer.