01
Lender Ladder
An AI practice tool built into a larger sales application for Home Lending Advisors and other frontline sales team members. Simulated calls let them practice relationship building and how to handle topics like net new money and objections.
My role
I was the main designer on a product team with a product manager and a developer. I worked closely with internal users at the Sales Academy, the training program for sales team members.
I designed the MVP, which included two agentic features: AI simulated calls for practice and an AI-generated feed.
The challenge
The operations team lacked scalable conversational intelligence. That led to inconsistent call quality and customer outcomes, and inefficiency through over-processing.
Managers also didn't have the capacity to cover every training need.
Approach
After speaking with the Sales Academy to understand their needs, I built a quick prototype using our internal LLM suite and VS Code. That got us to a proof of concept fast, and I kept iterating on the prototype from there.
The Sales Academy provided a prompt for each of the five main call categories: purchase, refinance, HELOC, real estate agent, and relationship-based pricing.
Preventing AI slop
I wrote clear prompts and programmed our internal design system into the prototype to prevent AI slop.
Screens
To respect my NDA, this isn't the real tool. It's a wireframe I generated as a basic visual of the experience, filled with sample data.
Outcome
Home Lending Advisor leaders believe the tool can reduce managers' workload and give team members a safe space to practice and grow.
Field Home Lending Advisors and central teams expect it to speed up engagement with learning materials and help people ramp up faster, so they can take on more customers.
What's next
The Sales Academy loved the idea of giving every call a more nuanced subject, by letting the AI create subcategories within each of the five main categories.