Onvero is a privacy-first platform pairing an AI legal assistant with a verified directory of non-profit immigration attorneys. How might we help undocumented immigrants feel safe enough to seek and access verified legal support?
Working alongside community organizations supporting undocumented immigrants across the Bay Area, I kept hearing the same story: people weren’t failing to find help because resources didn’t exist. They were failing because they couldn’t tell who was trustworthy.
“I’ve lived here most of my life, but I still feel like I’m walking on eggshells.”
Research participantWe designed primarily for Spanish-speaking community members with limited English and low digital literacy, the people most at risk of falling for a scam.
Who we designed for
Over one week, we interviewed 12 people: undocumented individuals, legal advocates, nonprofit leaders and community workers. Every session came back to trust.
Field interviews with community members and legal advocates
Community organizations and legal advocates we partnered with and who contributed to research
“Looking up legal help online felt like it could be a trap.”
Research participant“There are so many programs. It’s too much to figure out.”
Research participant“Google everything and sometimes I just give up.”
Research participant
Research synthesis, affinity mapping
Fear of scams, deportation, and exposure caused users to avoid seeking help entirely, even when support was available.
Legal resources existed across nonprofits and advocacy groups but there was no trusted place to discover them.
Language barriers, low digital literacy, and legal jargon created friction at every step of the process.
Crazy 8s + dot voting workshop
We used Crazy 8s and dot voting to explore ideas around trust, accessibility, legal guidance, and fraud prevention, then built our first version from the concepts that won.
Our first round of testing revealed a deeper issue than usability. Users didn't trust the platform.
They questioned where their information was stored, whether legal services were legitimate, and whether Onvero was safe enough to use at all.
Trust wasn't a feature users wanted. It was a prerequisite for engagement.
So before fixing navigation, we went back and rebuilt the experience around trust.
Users receive access through trusted nonprofits or existing members, making distribution itself the first trust signal.
Every lawyer or legal organization is validated through community partners before appearing in the product.
Privacy details appear upfront, so users understand how their information is protected before sharing anything.
Sketches exploring referral access, provider verification, privacy-first onboarding, and trust cues
Users encountered too many choices at once on the landing page. Features competed for attention.
Simplified to a single primary path and refined the UX writing to reduce cognitive load.
Large text blocks and inconsistent hierarchy made the interface feel intimidating. Users struggled to identify what was most important and needed to jump between pages, disrupting the flow.
Introduced clearer typography, stronger spacing, and progressive disclosure. Information became easier to scan and users moved through with more confidence.
Users questioned the AI assistant immediately. “Legal Buddy” felt generic and raised concerns about ownership, credibility, and data storage.
Rebranded as OnveroAI with transparent data handling explained upfront. Connecting the assistant directly to the verified platform increased trust.
Users didn’t believe services were truly verified. Without visible evidence, verification claims felt similar to the scams users were already trying to avoid.
Verification became transparent. Users can see which organization validated a provider and exactly why they can trust them before making any contact. Added a clear About Us section so users can trust us too.
The obstacle was twofold: users weren't sure OnveroAI was safe to confide in, and the AI and map felt like two disconnected tools. With feedback from Kyle Zantos, I clarified in the experience that OnveroAI is planned as a self-hosted AI, not tied to any third-party AI company, so user data would stay on the platform. I also connected the map directly into OnveroAI's suggested next steps, turning two tools into one path.
We designed for accessibility from the start: simple interactions, clear hierarchy, and colors checked for contrast and color-blind readability.
Contrast and color combinations checked for deuteranopia, protanopia and tritanopia, so text stays readable.
Users can speak to Onvero instead of typing, catering to low digital literacy or physical accessibility needs.
Starting with English & Spanish; with Spanish prioritized given the primary community served. Further language expansion planned post-launch.
All copy written at a sixth-grade reading level, with little to no legal jargon. The copy was tested with a Spanish-speaking participant with limited English.
Every screen evaluated against a single question: would an undocumented user feel safe using this?
Together we stand, together we navigate.
Privacy Before Participation
Before asking for anything, Onvero explains how user data is stored, who can see it and how it’s protected. Users get in through a referral from a verified community partner, not search or ads.
Transparency & Resolution
The About Us section shows the people and organizations behind Onvero. Users leave with a verified contact and a clear next step.
Trusted Access & Guided Support
OnveroAI then guides them through describing their legal situation in plain language, including voice input for low digital literacy.
Community Map & Verified Services
Onvero surfaces verified legal services showing exactly where each community organization is located. (Key Insight: Users trusted the service more when there was proof of a brick and mortar location.) Users see the evidence before making any contact.
Community Reviews
Users can review legal providers and share their experience. Hearing from people in similar situations gives the next person one more reason to trust a provider before reaching out.
After the trust redesign, more users finished the journey. One participant told us they felt safer knowing they weren’t doing it alone.
The same 10 people tested both rounds, so some of the improvement may come from familiarity with the task. We retested after the trust redesign to see whether it removed the barriers they hit the first time. Participant details are withheld to protect their privacy.
“I feel safer knowing I’m not doing this alone.”
User testing participant
For the final showcase, the team created a marketing video communicating the problem, the impact of legal scams, and how Onvero helps users find trustworthy support.
30-second conceptual story played during the final showcase at Wells Fargo HQ to illustrate the emotional impact of legal scams
In our first round of testing, users asked where their data went and whether the services were real before trying anything else. We had to fix trust before we could fix usability.
We started out designing for access to legal help. Research showed the real problem was safety: people could find help, they just didn’t trust it.
Our best design decisions came from nonprofits that already had users’ trust, like referral-only access through them. We couldn’t have built that trust on our own.
For this audience it also meant plain language, Spanish, voice input and feeling safe. That changed how our team thought about accessibility.
When users could see a verified contact and a clear next step, they were more willing to reach out for help they had been avoiding.
Users were sharing information tied to their immigration status, so we asked only for what we needed and explained why before asking.
Three areas of focus for the next phase of Onvero, each with early concept exploration.
Expanding beyond English and Spanish to Tagalog, Vietnamese, and Mandarin to serve a broader share of the undocumented community across California.
A fully voice-navigated mode for users with very low digital literacy, allowing them to access Onvero without any reading or typing requirements.
Growing the verified partner network beyond the Bay Area and building in-app case status notifications so users always know where they stand.