HOME

ROCKET MORTGAGE

HOME

ROCKET MORTGAGE

HOME

ROCKET MORTGAGE

From user trust to roadmap impact

I introduced interaction patterns to Rocket's AI assistant that made it to the product roadmap.

Context

TIMELINE

May 2025 - Aug 2025

INDUSTRY

Fintech

CONTRIBUTION

Product Design

COLLABORATORS

Conversational AI Designers, Product Designers

context

Rocket Mortgage's AI assistant guides first-time homebuyers through the mortgage process. At the post-offer stage, guidance inside the assistant was largely generic and disconnected from each client's specific situation.

contribution

I led the re-design for a personalized onboarding experience inside Rocket Assist, introducing task-based UI modules, card and carousel components and contextual prompt patterns. Some interaction patterns are now part of roadmap discussions.

“I'm buying a house; it's a lot of money, it should really be a guided journey”

A Rocket Mortgage client

the problem space

Where homebuyers start to feel lost.

After an offer is accepted, first-time homebuyers enter a 7 to 10 day window that requires completing inspections, reviewing documents and making decisions that directly affect their purchase. The guidance available inside Rocket Mortgage at this stage is is not personalized, with no account for the specifics of a client's home or nuances related to their mortgage journey.

72% of clients end up seeking help via chat, which drives agent call volume up at exactly the moment when support teams are already stretched.

"It took me a little bit of digging into the report to see what the appraisal was. It would be nice if there was just, like, appraisal came in at this amount."

A Rocket Mortgage client

SOLUTION 01

Task carousels that surface the right next step for each buyer's loan stage.

The most common question in the chat logs was some version of "what do I do next?" I explored moving the answer into the conversation itself, surfacing the most relevant next steps as task-based carousels inside the chat.

The onboarding flow introduced clients to Rocket Assist in the context of their specific loan stage, so the first thing they saw was directly relevant to where they were in the process.

client feedback

“...I like that it realized that the home has a pool so I would need a pool inspection done too and it’s giving relevant information...”

“...I like how the inspector cards are laid out and the information shown, feels super helpful. Everything is perfectly readable.”

Redesigned Onboarding Flow for Rocket Assist

INSPECTOR CARDS

DOCUMENT PARSER

Personalized recommendations of local inspectors

SOLUTION 02

Visual components that move clients forward without leaving the chat.

Text responses work well for simple questions. At the post-offer stage, clients are dealing with inspection reports, appraisal documents and multiple contacts across their loan team.

I introduced visual components inside the assistant: inspector recommendation cards personalized to the client's home, document download cards with highlighted insights and dynamic prompt buttons that surfaced the next logical action without requiring the client to ask.

client feedback

“I like that the inspector cards are easy to navigate and they are simple in design, I can see everything clearly”

“...I like that there is a direct link to the report and that there are highlights of the report. I like that I could download it instantly.”

SOLUTION 03

Direct human handoff when the stakes are high.

The conversation logs showed a consistent pattern: clients would reach a point where they needed reassurance from a person. I designed the escalation as a quick and transparent process.

Clients could transfer directly to their Purchase Specialist, see the estimated wait time upfront and schedule a callback if needed. The conversation history carried over so the specialist could see exactly what the client had already been through.

client feedback

“Overall, I am a big fan of the flow of connecting to a person when required and that the chat history is not deleted.”

“...I also like that it would notify me through the app when the person is available and I don’t have to save random numbers on my phone.”

Quick human handover and context preservation

IMPACT

Validating the experience WITH CLIENTS

92%

found tailored responses helpful

“I like that it [Rocket Assist] realized that the home has a pool so I would need a pool inspection done to, giving relevant info…”

75%

liked the inspector recommendations

liked inspector recommendations

“I like the inspector recommendations from my realtor, rather than those who might have paid for advertisement on this app."

90%

found transfer option reduced frustration

“I also really like that it gives me the option and the wait time to connect to a human being for stuff that is sensitive…”

BUSINESS IMPACT

The work moved beyond the internship into roadmap discussions.

Early and frequent conversations with the product team helped align the work around what was actually feasible and what could move fast. Several of the interaction patterns introduced are now part of ongoing roadmap discussions for the Rocket Mortgage mobile app experience.

"I am really excited about the direct escalation feature; this solves a real problem that the clients face. The current experience leaves them hanging in the void with just contact details. The proposed experience would be incredible!"

Digital Product Manager, RMO, Rocket Mortgage

TESTIMONIALS

From the people I worked with

"This was perhaps her most complex assignment, and Pri quickly mapped key friction points while collaborating with engineers and researchers. Her work helped influence product roadmap priorities."

DANA LEE

Director of Conversational AI Design & Digital Product Management

Rocket Mortgage

"Pri routinely sought out and addressed challenging issues, independently identified critical opportunities for improvement, and delivered results on par with a full-time associate designer."

AMANDA MATZENBACH

Conversational AI Design Manager & Mentor

Rocket Mortgage

REFLECTION

Not waiting to be pointed in the right direction is what opened the most useful conversations.

Working inside a mature design team for the first time, the difference was immediately noticeable. Watching senior designers think through downstream consequences before they became problems, reshaped how I approach my own work.

The moment that stuck with me most was presenting to the RMO product lead. I expected design feedback. Instead he broke each feature down to its smallest moving part. APIs, infrastructure constraints and engineering dependencies. It was the first time I understood that a single feature is really ten decisions that all need to align first.

Not waiting to be pointed in the right direction is what opened the most useful conversations. Seeking out engineers, designers and domain experts on my own got me into conversations I wasn't expected to be in and those conversations shaped the work more than anything else.

The question of how to keep a college advisor meaningfully in the loop — not just as a resource students could optionally contact, but as an active participant in the planning process, was something we acknowledged as future scope. In hindsight I wish we had considered it earlier. Not necessarily designed it, but used it to pressure-test some of our AI decisions along the way.

Presenting the project in a showcase to the larger Product Management & Design team @Rocket Mortgage.

Next project

PRODUCT DESIGN

2022

debug mode feature

I designed and shipped a debug tool that reduced testing time by 50%, for two distinct user groups.

PRIYAMWADA PANDEY

PLAY

RESUME

PRIYAMWADA PANDEY

Next project

PRODUCT DESIGN

2022

debug mode feature

I designed and shipped a debug tool that reduced testing time by 50%, for two distinct user groups.