SHOPPING FOR INSURANCE, WEB CRM
LEAD UX DESIGNER
2023 - 2025
:figma: FIGMA :figjam: FIGJAM :zeplin: ZEPLIN :storybook: STORYBOOK :tailwind: TAILWIND :salesforce: SALESFORCE :jira: JIRA :logrocket: LOGROCKET
The brokerage’s Salesforce CRM had grown brittle. Each new integration layered in another workflow mismatch. Agents were constantly re-entering data across tools, policy updates stalled in email threads, and clients received inconsistent documents. Growth slowed not because of lack of demand, but because operations were grinding under fragmented systems.
How can we help people get through complex insurance forms as
smoothly and as confidently as possible?
We built a proprietary web application that was split into two distinct client and enterprise experiences because agents and clients had conflicting priorities. Agents needed dense information tools for policy management, while clients wanted a stripped-down, fast path through forms. My focus was on streamlining agent workflows and making purchasing processes more intuitive. We also delivered a reusable library of web components, ensuring the business could continue to evolve the project.
Not Enough Overhead
As a lean design unit inside an enterprise, every decision had to prove impact. We anchored strategy around three levers: agent efficiency, client conversion, and technical scalability. This framing dictated scope: if a feature didn’t improve one of those levers, it didn’t ship.
Early discovery sessions with agents revealed that they were spending 30% of their time reconciling mismatched customer information between stickie notes, OneNotes, and what was actually recorded in Salesforce. In FigJam, we’d audit their workflows with the existing system, which exposed where absent or duplicate entries may occur. That insight directly informed our redesign of client onboarding.
Along with my Product Manager, we’d audit pages and annotate core touch points, trade offs, and interface friction. We’d re-work the dashboard to where instead of five linear, disjointed forms, it’d be a single dynamic form that allowed agents to jump between sections without the loss of data entry. This allowed them to structure their sales conversations in whichever order felt natural instead of them forcing the client through rigid inputs. Every design decision was tested and validated with agent focus groups, keeping our work focused more on their other regular needs.
To scale the application effectively, my teammate and I created a component library in Figma, later translated to Storybook, so core UI elements like form fields and tables could be reused across workflows without the need for intervention from our design unit.
We chose Tailwind as our front-end framework for its ability to white-label the system with native customization through root configuration, supporting B2B growth. Simple changes in color and branding assets would make the system bespoke to any of our business partners.
Clickable prototypes were tested with virtual agent focus groups. We observed where interaction cues failed or slowed them down, then refined workflows before development. This cyclical approach kept our work grounded in the agents day-to-day needs.
While agents wrestled with input inefficiencies, clients faced a different barrier: cognitive overload. The website presented them with long, jargon-heavy forms and only 4 policy options. Abandonment rates spiked midway through quoting because they didn’t understand what or why specific information was being asked. It was easier to call an agent, have them deal with recording information, and clients would have access to more coverage options than what the website forms supported.