Designing an identity verification experience that's easy to use

Regulated businesses have to know who their customers are, and identity verification is how they find out. Vaiie asked me to design the part of Identify that customers actually use.

Client: Vaiie
Categories: RegTech, Startup
Role: Persona Development, Customer Journey Mapping, Wireframing, UX Design, User Interface Design, 0 to 1 Design
Year: 2023

At a glance

Shipped white labelled, in each client's own brand, and working on mobile, tablet and desktop
Covers 254 countries and territories with more than 16,000 document templates
Went on to become a government-approved digital identity verification platform

The problem

Verification makes people stop, find an ID document and record a selfie before they can carry on. Every one of those steps is a place to drop off.

The journey also had to work anywhere. It might come during sign-up or by email weeks later. It might be embedded in a client's page, sit on its own URL or be built in through the API, on whatever device the customer had. On top of that, it had to carry each client's brand without hurting conversion.

The solution

Five Vaiie Identify persona slides fanned out. The front one profiles Greg, 36, a marketing manager applying for a car loan, with a familiarity rating for ID&V, background notes, and sticky notes covering his behaviours and frustrations

1. People, not number ranges

I built personas from research on past implementations, so the team had people to design for instead of number ranges to interpret. They covered everyone who touched the product, including back office staff setting up the theme, and included a set built with accessibility in mind.

Two user story mapping boards for Vaiie Identify, phase one and phase two. Both run across seven steps from welcome email to status, with persona-tagged story cards beneath each, and phase two adding four proposition cards setting out an assumption and a proposed solution

2. Letting users set the roadmap

I split the journey into steps, wrote user stories for each and traced them back to the personas. That settled what shipped first, which is usually the hardest conversation on a project.

Three features came out of it:

moving a session to another device
uploading an ID manually
guided capture with real-time feedback

Three stacked Vaiie Identify customer journey maps. The front one follows Greg, a marketing manager verifying his identity on a smartphone, across six stages from welcome to status, with rows for his actions, thoughts, an emotion line that dips negative during ID capture and the selfie, and colour-coded pain points and opportunities

A flow diagram of the Vaiie Identify verification journey, from the invitation email and device check through SMS one-time passcode, ID capture by webcam or manual upload, the selfie check and document upload, with branches for each failure and retry

3. Fixing the friction hotspots

The start: most people arrived unprepared, so the journey now says upfront what they'll need and how long it takes.
Equipment: no webcam or a failed device check now leads somewhere, not to a dead end.
Capture: the hardest part. I worked out that frame-line size plus video-player size sets how far away someone holds their phone, and those two numbers decide whether it feels comfortable or awkward.

Around thirty greyscale wireframes for the Vaiie Identify flow, covering phone verification and its error states, the device check, document type selection, photo guidance, selfie video recording and supporting document upload

4. Designing the unhappy paths too

I wireframed the whole journey, not just the main screens: a one-time password that fails, a device that can't cope, a check that comes back negative. Every heading, line of help text and button label was written to move someone one step further.

The Vaiie Identify liveness check on a phone held up in a bright room: a camera preview with an oval face guide, instructions to press Record and follow the on-screen prompts, and a Record button

5. Testing what the wireframes couldn't show

The selfie included a liveness check, where people read a random number aloud. Shown "33-44-11", most said "three three four four one one", but the check needed "thirty-three forty-four eleven". The technology fixed the spoken test, so the only lever was how the number was presented.

The camera also wasn't mirrored, so anyone adjusting their position moved the wrong way.

Four phones showing the first four steps of the Vaiie Identify verification flow: entering an SMS security code, framing the photo page of a passport, positioning a face within an oval to record a liveness check, and uploading supporting documents

Outcomes

Vaiie Identify shipped working on mobile and white labelled, covering 254 countries and territories with more than 16,000 document templates. It went on to become a government-approved digital identity verification platform.

Jon branded our entire suite of regulatory technology tools that sit at the heart of the business.
He brought a real moment of clarity during a period of complexity - from brand identity and product design through to working prototypes, and from managing agencies to working directly with clients, he made it all look effortless.