Working backward for a smarter insights platform
As the founding designer on Yogi’s platform team, I had the opportunity to lay the foundation for a scalable B2B product that could make customer feedback not only readable — but actionable. Moreover, the challenge of growing the userbase to reach executives & novice data-facing roles.
In an early-stage environment, priorities shift quickly, so I focused on establishing a clear design direction and working rhythm for the team. My responsibilities ranged from leading a full platform redesign to launching an automated insights experience, hosting internal hackathons, and prototyping future-facing AI features.
Problem
Yogi was poised to start growing quickly — with large CPG clients demanding speed, clarity, and reliability in their product insight workflows. But the platform lacked visual hierarchy, didn’t support evolving use cases, and couldn’t scale across verticals without adding UX debt.
Platform Redesign
We kicked things off with a series of user interviews to understand how insights teams and brand managers currently navigated the product. I mapped out core journeys and built interactive prototypes that restructured the navigation, dashboard hierarchy, and insight presentation.
Our goals:
- Reduce time-to-insight
- Elevate key metrics and trends
- Modernize the platform without alienating power users
- Reduce time-to-insight
- Elevate key metrics and trends
- Modernize the platform without alienating power users
Launching LLM Insights
The redesign was great for sales, having a record quarter. We want to scale within orgs that only had a few users so we did some research and launched a new feature that delivered real-time trend recognition powered by natural language processing and sentiment clustering. I worked closely with the data insights team to turn models into moments — designing digestible, contextual summaries with clear calls to action.
after user testing, this project took a few iterations to nail customer intrigue, which involved fine-tuning the AI to Generate Reports that felt more like the news!
Internal Hackathons
To keep the team creatively sharp, I led design in a series of hackathons across design, product, and engineering. My role was to inspire new thinking, prototype rapidly, and pitch ideas that could ship — two of which were prioritized for roadmap inclusion.
This exposed 3 big projects and one of them became our Chat Research + Deep Thinking LLMs, and avoid the repetitive chatbot interactions users have seen.
One Hackathon project customers loved immediately was this AI-Assisted Deck Creator that uses AI to help you make databacked claims!
User Testing for a new Onboarding
Throughout the process, I ran dozens of 1:1 sessions with client teams and internal stakeholders. We validated flows using moderated testing on live accounts, implemented lightweight intercepts for fast feedback, and integrated usage data to refine onboarding paths.
A key insight: researchers valued flexibility, but needed guardrails — leading us to implement modes and toggles for advanced vs. guided workflows.
Final Takeaway
Designing the platform from zero gave me a chance to build trust through utility. Keep making new perspectives of the same data is a challenge worth taking. Whether it was simplifying navigation, launching ML-powered insights, or prototyping AI tools that felt like actual coworkers — everything pointed back to one mission: make it easier for teams to hear what customers are really saying, and act fast. CPG as a market is getting smaller and smaller every year, so acquisitions and consolidation is underway.
Bonus!!
I got to create some AI Video for us, and we used it for Ads, great performances!