Case study / Compass, formerly Anywhere Real Estate

Keysign

I turned an engineering-led prototype into a clear, scalable tool that helps agents handle the new paperwork that came with the 2025 NAR settlement.

  • 0 to 1
  • End-to-end
  • B2B
Keysign welcome screen on mobile: create and sign agreements in 4 steps
Role
Senior Product Designer, leading the end-to-end redesign
Team
2 Sr. Product Managers, Director of Engineering, Sr. Software Engineer, design leadership, accessibility and a UX writer
Timeline
3 months to get it in front of real agents
Scope
User research, UX design, prototyping, storytelling
The problem

After the NAR settlement, every showing needed paperwork, legal disclosures and documented consent. Agents were printing templates, filling them out, scanning them and waiting on a signature before they could show a home.

My approach

Engineering had already built a mobile prototype. Instead of starting over, I made what was there clearer and more scalable, with small changes that didn't need back-end work.

Results
70%
less time on task
20 – 30 min down to 5 – 10
83%
fewer errors
67%
higher adoption
Now in the flagship agent software
180k
agents 
01 / Context

Scrappy prototype to scalable CX

Engineering built a mobile-first prototype where agents could complete templates and send agreements to buyers for signature through DocuSign. UX joined later. The CTO wanted to test it with real agents within three months, and the design teams were at capacity, so I took ownership of the design and led the effort to launch while still supporting my teams.

Paper contract templates next to an early sketch of the mobile flow
How agents handled paperwork before
Problem statement

How might agents draft paperwork more efficiently and clearly, without rebuilding the app from scratch?

02 / Discovery

Time-cruuuunch

The prototype mapped every field from the agents' DocuSign templates into one mobile form. In real use, an agent working across several metro areas found it overwhelming. It also needed to pass accessibility and content reviews and move from Material UI to our design system.

Engineering had built the foundation on a tight timeline, so any design change had to protect delivery dates. Research time was limited, so I aligned with product on user insights and business goals, studied the legal and compliance requirements, and worked with design leadership and accessibility partners.

Diagram of the existing architecture and agreement flow
Understanding the architecture before changing it
03 / Strategy

Workflows for busy agents

Minimize cognitive load

Agents fill out forms in their cars or in front of a house with buyers waiting.

Stay predictable across jurisdictions

An agent might show homes in Seattle, Bellevue and Tacoma, each with its own templates and rules.

Build trust through clarity

The tool should help the agent's workflow, not get in its way.

Respect engineering constraints

There was no time to rebuild, so I focused effort where it mattered most.

Set up long-term integration

If it worked, it would become a core feature of the agent platform in development.

04 / Constraints

Working within tough constraints

The trickiest part was working within the technical constraints. I looked for changes that would make a real difference without adding back-end work or pushing the timeline.

Templates and agreements, separated

Splitting Templates and Agreements into tabs was a low-effort change with a big usability gain.

Keysign home screen before and after splitting templates and agreements into tabs

The second buyer

In the original, adding a second buyer meant finding a grayed-out icon at the bottom of the list, opening it and switching the form on with a toggle. Not intuitive.

Original second buyer pattern with a hidden toggle
The original pattern

I explored asking upfront how many buyers there were, an Add button and dynamic fields. Engineering needed back-end changes for all of them, and there wasn't time.

Wireframe explorations for adding a second buyer
Explorations that needed back-end work

So I stepped back and proposed a copy change: (Optional) next to the second buyer's section. No logic changes, no new components, no back-end work. Product and engineering loved it, and it worked.

Second buyer section labeled Optional in the final design
The fix
05 / Final design

Ready for launch

The final experience kept engineering's architecture but changed how agents move through it. Known information pre-fills, and better filters make agreements easier to find.

Filters

Engineering added a way to take or upload a photo of the buyer's ID to fill in fields, and the UX writer made sure every instruction was clear. Because agents consent on the buyer's behalf, we wrote that language together so it would pass legal review.

Agreement flow before and after the redesign
Before and after
The original flow
The redesigned flow

After launch

Agents review, approve and send the agreement to the buyer through DocuSign. Based on their feedback, and with engineering on board, agents can now sign directly in the app instead of detouring into DocuSign.

Sign in app
06 / Impact

The impact

Results
70%
less time on task
20 – 30 min down to 5 – 10
83%
fewer errors
67%
higher adoption
Now in the flagship agent software
180k
agents it rolled out to

Agents reported lower cognitive load, with fewer errors and less back-and-forth with buyers.

Leadership saw the value right away, and Keysign shipped as a feature of the flagship agent software. For the UX team, it built stronger trust between engineering and design, and a validated path for scaling into agent software.

07 / Learnings

What I learned

The biggest lesson wasn't about workflows or templates. Great experiences are possible under heavy constraints when clarity, empathy and creativity lead. Being both a design leader and a hands-on problem solver here was challenging and rewarding, and bringing UX in early builds credibility and trust with partners.

Confidential information has been omitted or blurred to comply with my NDA.