KYB — Institutional Verification Redesign

Reframed an “add a selector” request into a root-cause redesign of document submission — for the institutional segment that drives 30–70% of exchange trading volume.

01 · Problem

The why

The request was cosmetic. The real problem was comprehension.

KYB is how institutional clients verify their company to trade. Institutional clients drive 30–70% of exchange volume — every drop-off here is lost revenue in the highest-value segment.

A user research report flagged a low KYB conversion rate. A feature request landed on my desk to “fix” it.

The request I was handed

“Add a company-type selector so users only see the fields relevant to them.”

A reasonable-sounding feature request.

What I found — traced it back to the user research report myself.
Why the request missed — heaviest drop-off sat on documents nearly every company type submits; filtering by company type wouldn’t touch the real bottleneck.
Two failure modes — users submitted without reading requirements and got rejected; others stalled at document upload and gave up.
Bottom line — of everyone who submitted documents, only 41% reached approved KYB.

—

Company info submitted — 169

89%

Contact person submitted — 151

75%

Documents submitted — 114

41%

KYB approved — 47

Nov KYB self-submission funnel · % = conversion from the step above. Starts at company-info submission, excluding pre-intent homepage traffic the report flagged as non-institutional noise.

02 · Data-driven insight

Prove the why, then pinpoint the where

82% resubmitted — a clarity problem, not a motivation problem.

What the data showed

Users weren’t abandoning out of disinterest — they were failing, resubmitting, and stalling. These three numbers set the baseline.

41%document submissions passed review

82%resubmission rate — primary friction signal

~40%took more than 5 days to complete

Why — the failure chain

From the research report and internal interviews — not a motivation drop-off, a comprehension loop.

Wall-of-text requirements → wrong documents submitted → rejection → abandonment

Where — two modules drove ~90% of failures

CBR / PoA — wrong document

67%

Shareholder register (>25%)

24%

Other modules

Share of document-submission failures by module (Nov).

What to fix — two precise targets

① CBR vs PoA · 67% wrong

Board vs no board

The split — with vs without a board of directors, two different legal documents apply.

The failure — buried in a wall of text; users never noticed and uploaded the wrong one.

② Shareholders > 25% · 24% wrong

Ownership thresholds

The rule — shareholders owning >25% trigger additional documents; further conditions add even more.

The failure — users couldn’t track cascading requirements from static text.

So what

Redesign focus

Takeaway — fix these two modules and the redesign has a clear, measurable focus.

Previous design — the two document modules where most submissions failed

Previous design — key distinctions buried in dense text at exactly these two modules.

03 · Approach

Benchmark the pattern, then decide what to keep

Competitor research was an input — alternatives were my decision.

What leading platforms do

Three patterns recurred across Binance, OKX, Coinbase and Kraken — inputs to my solution, not the answer.

Pre-select entity type — route users by institution type upfront.
Document checklist — show what to prepare before they start.
Plain-language instructions — readable requirements at each step.

Three recurring patterns across leading exchanges — pre-select entity type, upfront document checklist, plain-language instructions

Benchmarking institutional verification flows across leading exchanges.

Alternatives I considered — and how each earned its place

Option A · kept for scale

Company-type selector

Today’s bottleneck — doesn’t fix comprehension failures.
At scale — as EU and AU regulation expands, document requirements diverge by region; the selector routes the right set per market.

Option B · supporting role

Help text

Users already skip long text, so I used it only for basic concept explanation — never as the fix for complex requirements.

Option C · core fix

Guided required choice

Routes users to the right document with the lowest cognitive load — no hard reading required. This carried the redesign.

04 · Solution

The build

Turn reading into a guided, unskippable decision.

1. Guided selection — the board / no-board distinction became a required choice that routes users to the correct document — the step that was 67% wrong before.

Guided selection replacing text instructions

2. Visual hierarchy — requirements are scannable instead of buried — directly targeting the 24%-wrong shareholder step.

Restructured requirements with clear visual hierarchy

3. Additional upload section — raises approval chances when compliance requests further documents after the first submission.

Additional upload section for supplementary documents

05 · Edge cases

The messy parts

The states a happy-path demo hides.

1 / 3

More materials required

High-risk applicants get prompted for additional documents after submission, each with a clear reason — so the request doesn’t feel like a dead end.

2 / 3

Verification failed

Rejections show specific, fixable reasons so users can debug and resubmit correctly — instead of guessing and abandoning.

3 / 3

Continue later

Institutional clients need time to gather documents, so progress is saved and the flow is resumable from where they left off.

06 · Final design

The flow

Select entity type → document checklist → guided submission.

Select entity type — jurisdiction-specific routing.
Document checklist — exactly what to prepare before starting.
Guided submission — company info, contact person, required documents, and an additional-document step — with the right-document decision made at each fork.

End-to-end KYB flow — select institution type, document checklist, steps overview, then four guided submission steps

Full journey — entry, preparation, and the four submission steps.

Key screens — company info, contact person, and required documents with guided upload modules

Key screens — the guided form steps where comprehension decisions happen.

Try it yourself

Interactive prototype — click through the guided flow. Open in Figma ↗

07 · Impact

My contribution

What I owned — and why it’s better than before.

My role

Reframed the brief — from cosmetic selector to root-cause redesign.
Dug into source data — established baseline and success metrics.
Designed IA, flow & UI — end-to-end, with 1 PM and front-/back-end engineers.

Better than the previous version

Before

A wall of text

Static requirements users skipped → wrong submissions → rejection → drop-off, with no clear way to recover.

After

A guided decision

A structured choice that routes users to the right documents with no added reading — plus a rejection loop they can actually debug.

Outcome & measurement

41% Document submissions approved — baseline

82% Resubmission rate — baseline

10–15% Projected lift vs. comparable KYC optimization

Target — reduce wrong submissions and completion time.
Early read — first-time pass rate edged up from 41% to 42% — a small, real lift, well short of the projected 10–15%. Engineering capacity meant only part of the core capability shipped.
Tracking — built the measurement framework so the full impact can be captured as the rest ships.

Scaled across markets

The same guided-verification pattern became the base for the EU and AU institutional flows — re-applied to materially harder compliance logic: multi-layer entity structures, beneficial-ownership thresholds, and jurisdiction-specific source-of-funds rules (trust deeds, SMSF lookups).

Business outcome

The EU and AU KYB/KYC flows supported the company’s licensing submissions under local regulations.

Design that helped unlock regulated market access — not just a flow improvement.

KYB EU and KYB AU — the same guided-verification pattern adapted to local regulatory requirements

KYB EU and AU — the same pattern extended to each market’s entity types and document rules. Full screens available on request.

08 · Hindsight

Self-awareness

What launch taught me — and would change next.

After launch, first-time pass rate rose only marginally (41% → 42%) — well short of the projected lift. The more useful question was why it wasn’t larger, so I sat down with customer support and the compliance team.

The real gap: hidden requirements. The interface never showed the detailed document rules — so users failed checks they were never told about. A business certificate needed a certified-true-copy stamp, not just the original. Some documents needed a manual signature that users almost always skipped.
What I’d change: make every rule visible at the point of upload. Guiding users to the right document isn’t enough — I’d also show what makes it valid (stamp required, signature needed) right where they upload, not buried elsewhere.
Ship the highest-impact piece first. I’d scope the rollout with engineering earlier, so guided routing lands first and the metric can move sooner — instead of a partial release.

Institutional Verification Redesign (KYB)