Modernizing Remote Exam Check-In for Global Credentialing

Remote proctored exam check-in is not onboarding. It's a 30-minute pre-flight where a candidate has to verify their identity, capture their ID, complete a 360° room scan, demonstrate a clean testing environment, agree to strict rules, and pass technical readiness checks - all under time pressure, often in a language that isn't their first, and knowing that any failure could mean missing a high-stakes certification window.

Behind the scenes, the platform has to validate hardware, network strength, browser lockdown, identity match against the booking, and environment compliance - while protecting candidate privacy, satisfying regulatory bodies, and maintaining exam integrity at global scale.

I led product design for this end-to-end check-in journey across desktop and mobile, evolving it from a brittle Flash-era acquisition into the cross-device foundation that supports a platform now delivering nearly 21 million exams annually.

The starting point was a check-in tool acquired from a third-party vendor in 2016. It was built in Flash - a runtime that was actively being phased out across browsers - and locked to a fixed 468×600 pixel viewport that ignored the realities of modern devices, screen sizes, and accessibility expectations.

That created cascading operational problems:

• Device dead-ends. Candidates on tablets, large monitors, or non-standard configurations hit broken layouts and unclear next steps.
• Brittle handoffs. Switching between desktop and phone (required for selfie capture and ID photos) was confusing - candidates didn't know which device to use, when, or what was happening on the other screen.
• Failure under pressure. When something went wrong, the candidate often didn't know whether it was their fault, their device, or the platform - and recovery paths were unclear.
• Browser-deprecation risk. The Flash sunset was an existential business risk. Modernization wasn't optional.

The discovery work centered on understanding what candidates were actually experiencing - not the idealized happy path but the messy reality of taking a high-stakes exam from your kitchen table.

What emerged from journey mapping, support-ticket analysis, and operational feedback:

• The 30-minute window is psychologically loaded. Candidates often arrive early, anxious, and worried that any technical hiccup will disqualify them.
• Cross-device is unfamiliar. Most candidates aren't expecting to use their phone alongside their computer. The handoff has to be obvious, fast, and clearly bidirectional.
• Identity verification is intimate and high-stakes. Candidates are asked to photograph themselves, their government ID, and their workspace. Small confusions create big failures.
• The environment scan is unique. A 360° room scan via webcam or phone is unlike any other consumer experience. Candidates need clear, calm, step-by-step guidance.
• Recovery has to be possible. If a check fails (bad lighting, an item visible on the desk, a network drop), the candidate needs to know what's wrong, how to fix it, and that they still have time to test.

The strategic frame I held to throughout the multi-year effort:

• Continuity over completeness. Don't try to ship every modern UI pattern at once. Migrate the platform safely, in phases, without ever degrading exam integrity or the candidate experience.
• Phone as assistant, not silo. The desktop is the test surface. The phone exists to help - selfie capture, ID photo, room scan - and then hand back. The two devices should feel like one coordinated system.
• Reduce uncertainty before reducing time. The check-in isn't long because steps take time; it's long because each step creates uncertainty. Reducing anxiety improves throughput more than shortening flows.
• Design for the failure path. The first-time candidate who needs help is the design target. The repeat candidate who breezes through is a happy side effect.
• Design system over component design. Every new pattern had to plug into the broader 50+ application ecosystem - shared tokens, reusable interaction patterns, consistent accessibility behavior.

1. QR-first device handoff. Replaced the original manual-URL-entry pattern with a scannable QR code that bridged candidates from desktop to phone in seconds. Eliminated typo-driven failures and reduced time-to-mobile by orders of magnitude.

2. Persistent Mobile Bridge. Built a coordinated state model where desktop and phone always agreed on "where we are." Candidates on the phone could see what step they were completing for the desktop session; candidates on the desktop could see live progress from the phone.

3. Photo capture flows tuned for the room. Designed selfie and ID photo flows that worked under poor lighting, awkward angles, and one-handed phone use - with clear retake paths and progressive disclosure of what the system needed.

4. Environment scan as guided tour, not test. Reframed the 360° room scan from a compliance gate into a guided experience: step-by-step prompts, real-time framing feedback, and recovery guidance when something needed to move or change.

5. Hardware-agnostic flow logic. Designed the check-in to gracefully handle device variability - different camera resolutions, browser quirks, network strength, and operating system behaviors - without surfacing technical complexity to the candidate.

6. Trust-critical state communication. Built clear, calm UI states for every moment of uncertainty: "we're checking your ID," "your environment looks good," "we need you to try again - here's why." Avoided both alarm and false confidence.

Phased modernization vs. a single rebuild. Pushing for a complete rewrite would have been faster on paper but catastrophic in practice - the platform was actively delivering exams to millions of candidates and couldn't go dark. I designed a phased migration path that let engineering modernize underlying services without ever stopping the world. The cost was years of carrying both old and new patterns in parallel; the benefit was zero candidate-facing disruption.

Cross-device complexity vs. desktop-only simplicity. The simpler design would have kept everything on desktop. But identity verification, room scans, and ID capture work demonstrably better on a phone the candidate already has in hand. I leaned into cross-device as a feature, not a workaround - and invested heavily in making the handoff feel intentional.

Strict rules vs. forgiving recovery. Exam integrity demands strict rules: no other people, no extra screens, no notes. But strict rules without recovery paths convert anxious candidates into failed sessions. I designed recovery flows for every failure mode - but kept the rules themselves non-negotiable.

• 300% user volume growth over the period of design stewardship, with the platform absorbing major surges (including the global shift to remote testing) without breaking the candidate experience.
• 25% reduction in check-in failure rates during the most aggressive period of platform acceleration, when a 24-month roadmap was compressed into 6 months.
• 13M+ online proctored exams delivered cumulatively on the modernized platform, with public scale signals showing the platform now operates in 180+ countries and territories, in 47 languages.
• A foundation for what came next. The cross-device, system-aware design pattern I established became the platform groundwork for AI-assisted verification, computer-vision environment scanning, biometric identity confidence, and the multi-tier proctoring model (live, hybrid, review-based) the platform now offers.

The most enduring lesson from this work is that modernization is stewardship, not innovation theater. The senior design contribution wasn't a single beautiful screen - it was the patience to design migration paths, the discipline to keep candidate trust intact across architectural change, and the systems thinking to ensure each new pattern strengthened the broader platform.

If I were starting this work today, I'd push earlier on three things: AI-assisted readiness coaching (helping candidates prepare their environment before the 30-minute window starts), continuous identity confidence (lightweight checks throughout the session rather than concentrated at check-in), and richer multilingual support across the most error-prone moments - because in a global platform, the candidate having the worst experience is rarely the one whose first language matches the UI.