Enter password to view case study
THE PROBLEM AT SCALE
Of seller time lost
AE's and SA's on research and admin - not customers
Field reps in scope
AE's and SA's as primary users across AMER, EMEA, APAC
Per strategic deal
Lost to artifact creation - POC docs, emails, briefs
CONTEXT
The problem Databricks was trying to solve
Databricks' GTM field teams were spending 40–60% of their time on work that could be automated — hunting for account intel across Slack, email, and Zoom; manually updating Salesforce after every customer call; and building sales artifacts from scratch for every deal.
"We spend increasingly more time navigating complexity, distilling information, and performing tedious work that could be automated. We are wasting so much time."
— L7 Solutions Architect, Databricks · 4 years tenure · Internal MVP review, Oct 2025
MY ROLE
I led UX end-to-end, collaborating with PM and engineering
I was brought in to lead UX for SAI, starting with evaluating early design concepts and evolving them into a fully production-ready experience that ultimately expanded into a ground-up redesign.
As the sole designer, I was deeply embedded with PM and engineering in daily standups and design reviews, translating requirements and critical user journeys into designs that could move directly into development.
I regularly presented to Director-level stakeholders, facilitating structured design reviews, iterating based on feedback, and clearly articulating design decisions in the context of speed and business impact.
The team operated at a rapid pace, with tight feedback loops and continuous collaboration enabling designs to progress from concept to build in days rather than weeks.
USER RESEARCH
Research - informed design across the full product lifecycle
When I joined, research was already in progress and a prototype had been tested with real sellers, giving the team a clear view of key friction points. I supplemented this by conducting additional user interviews to close gaps and validate specific design decisions.
Rather than starting from scratch, I built on the existing foundation—using prototype feedback as my design input, leveraging PM-captured user insights as a continuous signal, and running stakeholder reviews to validate major decisions before engineering handoff.
01
Field discovery & workflow mapping
AE's • SA's
Through interviews with field reps across multiple business units, the team identified key workflow bottlenecks and pain points. The findings guided product strategy and feature prioritization, ensuring development efforts were aligned with real user needs.
02
Prototype testing & stakeholder feedback
Lovable prototype
Initial prototype shared with users to validate the core concept. That feedback — what resonated, what confused, what was missing — became the brief I inherited and designed against.
03
Iterative Figma design reviews
Stakeholder walkthroughs · Ongoing
As I produced Figma designs, I ran review sessions with GTM stakeholders and business owners. Their feedback directly shaped interaction patterns, content hierarchy, and edge case handling — closing the loop between research insights and final design decisions.
Some key findings and design decisions
Finding
Design Decision
Sellers were required to update Salesforce on a fixed weekly cadence regardless of whether meaningful activity had occurred. Meeting notes lived across Zoom transcripts, email threads, and personal notebooks — scattered with no connection to the UCO that needed updating. Reconstructing what happened, finding the right note, and translating it into Salesforce field language was a memory exercise that happened every week, for every account.
The problem wasn't that sellers didn't want to update Salesforce — it was that updating required too much mental reconstruction. I addressed this by designing the agent to do the gathering for them. Every alert in My Actions arrives pre-drafted, source-cited, and editable — sellers are reviewing a proposal, not building one from scratch
Sellers distrusted AI outputs without visible sourcing. They would dismiss recommendations even when correct if they couldn't immediately understand why the system was suggesting an action.
Source traceability became non-negotiable on every AI surface. Every recommendation, alert, and draft update shows the exact source snippet — Zoom excerpt, Slack message, or email line — that triggered it. No black boxes.
Sellers weren't just struggling with CRM hygiene — they were also spending significant time building demos, finding relevant slides, and developing talk tracks before every customer meeting. Pre-meeting prep was a top time sink.
Meeting Briefs were added as a core feature — auto-generating structured prep materials 24–48 hours before external meetings — directly in response to this finding. Five brief variants cover executive, technical, discovery, partner, and standard meetings.
PROTOTYPING
Research - informed design across the full product lifecycle
My prototyping approach started with a critical audit of the existing Lovable prototype — identifying what worked, what didn't, and what was missing entirely. I then redesigned the full product in Figma, working directly to production-quality mockups rather than wireframing first. Clickable Figma prototypes were used throughout to run design reviews with stakeholders before any design went to engineering.
Lovable prototype - What I inherited
•
Fast to build and good for concept validation with early stakeholders
•
Inconsistent interaction patterns across different surfaces
•
Missing edge cases, error states, and loading behaviors
•
No design system — one-off components throughout
What I delivered
•
Using Databricks design system — reusable components, tokens, and patterns across every surface
•
Full state coverage — default, loading, empty, error, and success for every feature
•
Interactive clickable prototype for stakeholder reviews and user walkthroughs
•
Engineering-ready specs with annotations for complex interactions
USER INTERFACE
REFLECTION
What this project taught me
Stakeholder communication is part of the design
Presenting to VP and Director-level leadership regularly taught me that how you frame a design decision matters as much as the decision itself. Clarity, brevity, and business relevance — not just craft — are what move designs forward in fast-moving orgs.
The problem brief is rarely the real problem
The initial ask was to improve an existing prototype. The real problem was that sellers had no reliable way to keep Salesforce current or identify strategic opportunities at scale. Getting to that framing was the most valuable design work I did on this project.
Pace and quality can coexist with the right team
Moving from concept to handoff in days forced discipline around what mattered most in each iteration. Tight daily loops with PM and engineering meant decisions landed fast — and the product was better for it.