Senior Product Designer · Take-home

The Arta Card
Extending the wealth relationship into everyday liquidity.

00 / 19
Opening · 01 · My read of the brief

The brief became more interesting once the product model was clear

01 · The brief

Launch the Arta Card

Design a smooth application experience, an AI relationship across the card lifecycle, and a model that can extend to partner brands.

02 · The clarification

A portfolio-secured charge card

Existing Arta members already have a verified relationship, and eligible custodied assets can inform prequalification and the recommended limit.

03 · The reframe

This isn't a cold-start credit relationship

Arta already knows who the member is, holds part of their financial context, and may hold the assets supporting the card.

So the opportunity isn't to make a traditional card application prettier.
It's to design credit around an existing wealth relationship.

01 / 19
Opening · 02 · Assumptions

The boundaries I designed within

Audience

Existing, KYC'd Arta members

I designed for members with an established Arta relationship rather than a cold-acquisition journey.

Market

Singapore-first working assumption

I used Singapore as the launch market and the HSBC Singapore form as the reference for information and disclosure scope. Exact regulatory requirements would be validated with Legal and Compliance.

Product

Portfolio-secured charge card

Eligible custodied assets and the existing Arta relationship can inform prequalification and the recommended limit. Portfolio use is explicit and member-controlled.

Experience

Mobile-first, inside Arta

The launch experience lives inside the existing Arta app. I designed for both instant decisions and cases that require additional review.

02 / 19
Opening · 03 · Audience

I grounded the design in Arta's original customer thesis

01 · Who Arta set out to serve

Successful professionals in the wealth gap

Arta's founders described an underserved group of professionals who are doing well, but are not yet at the level where a private bank or family office serves them well.

NOT MASS MARKET → NOT ULTRA-HNW →
EMERGING / ESTABLISHED WEALTH

02 · Where they come from

Digital-native professionals hitting their financial stride

Arta specifically called out professionals in their 30s and 40s across technology, consulting and finance.

TECHCONSULTINGFINANCE

03 · What makes their wealth different

Income alone may not tell the whole story

Arta has also described members with significant employer-stock exposure and concentrated positions — particularly in technology.

SALARY + COMPANY EQUITY + INVESTED ASSETS

FINANCIAL CAPACITY ≠ PAYCHECK ALONE

Design implication

I shouldn't design Arta Card for a generic credit-card applicant.
I should design for members whose wealth may sit in assets, equity and investments — not only salary.

03 / 19
Opening · 04 · Design targets

From Arta's audience thesis to two design archetypes

Scope choice: I considered a broader HENRY audience, but designed only one primary journey and one meaningful stress test to preserve depth.

One represents the core audience I found in Arta's public positioning. The other is deliberately constructed to stress-test where that model becomes difficult.

ARTA'S STATED AUDIENCE — SUCCESSFUL PROFESSIONALS BUILDING MEANINGFUL WEALTH
↓ DESIGN TRANSLATION ↓
CORE REALITY · EQUITY-RICH OPERATOR → REMOVE FRICTION EDGE REALITY · LUMPY-INCOME FOUNDER → TEST THE LIMITS
Arta Card · Archetype01 / 02
Research-grounded primary archetype

The Equity-Rich Operator

A successful professional with significant compensation and invested wealth, including employer equity · demoed as Alexandra
IncomeHigh compensation, partly equity-based
WealthMeaningful assets already held with Arta
Portfolio patternPotentially concentrated in employer equity
Core tensionLiquidity without unnecessarily disrupting long-term investments
What this tests

How much friction can Arta remove because it already knows the member and their assets?

"Give me access to liquidity without making me manage another financial system."
Reuse existing contextLiquidity without liquidationClear terms & control
Arta Card · Archetype02 / 02
Hypothesis-led stress test

The Founder with Lumpy Income

Asset-rich, but with irregular income that conventional salary-led application patterns may struggle to represent · demoed as Dev
Income patternIrregular and event-driven
Salary evidenceMay not have conventional payslips
WealthSubstantial but potentially illiquid or concentrated
Core tensionTraditional income signals may underrepresent financial capacity
What this tests

Where can portfolio context complement traditional evidence — and where must Arta fall back to review?

"Understand my financial reality without forcing it into a salaried-worker form."
Portfolio-aware assessmentExplain the exceptionTransparent review
BOTH ARCHETYPES ARE SWITCHABLE IN THE PROTOTYPE ↗
04 / 19
Opening · 05 · The design question

One design question guided the work

The tension is the point: reduce friction — but never by reducing transparency. Three principles operationalize it, and each one owns a section of this deck.

How might we use Arta's existing member and portfolio relationship to remove unnecessary friction from credit — without removing transparency or control?

01 · Reuse what Arta knows → Flow 1

Don't ask twice

Reuse verified member and portfolio context wherever it is safe and appropriate.

02 · Protect member agency → Flow 2

Explain before acting

Make consequential decisions understandable, and keep the member in control.

03 · Design for the platform → White-label

Keep trust in the core

Let partners own the brand expression while Arta protects decisioning, compliance, security and AI guardrails.

05 / 19
Opening · 06 · Risk-first process + deliberate cuts

I spent the 5 hours where the product risk was highest

30M 45M 2 HRS · THE HIGHEST-RISK FLOW 1 HR 45M 30M
01
30 MIN

Frame the product

Clarifications · assumptions · success criteria

02
45 MIN

Audit the existing experience

Arta product language · HSBC disclosure scope · system gaps

03
2 HRS

Solve the highest-risk flow

Application · disclosures · decision states

04
1 HR

Define the AI relationship

Lifecycle map · one deep moment · guardrails

05
45 MIN

Prove it can scale

White label · component & token model

06
30 MIN

Edit the story

Tradeoffs · cuts · presentation

Deliberately not designed

Rewards catalogue · full transaction management · disputes · desktop/web · supplementary-card flows

06 / 19
Flow 1 · The application — Principle 1 · Don't ask twice

Confirm what Arta knows. Collect only what credit needs.

  • Arta already KYC'd this member — the flow confirms what Arta knows and asks only what credit still needs. Most of the application is confirmation; additional verification appears only when needed.
  • One application, adaptive friction: the rules stay consistent — the two members branch only where the evidence differs, and at the decision.
  • Honesty is the premium experience: key economics appear before commitment; full terms appear before submission.
Entry — pre-qualified card in the grid Confirm Complete Secure Settle Commit · Key Terms Decision · 5 governed states

Every screen ahead is the live prototype, embedded.

OPEN THE LIVE PROTOTYPE ↗
07 / 19
Flow 1 · The application

Entry & confirming what Arta knows

The offer, in the portfolio gridPre-qualified from what Arta already holds — key terms stated before the funnel
What to expectFour steps, mostly confirmation — with the trust block up front
Confirm, don’t collectThe verified profile shown as review cards, not input fields
08 / 19
Flow 1 · The application

Income is sensitive. The explanation shouldn't be optional.

Arta knows who you are. Singpass proves what you earn. SG-market pattern — exact Myinfo attributes and consent scope would be confirmed with Compliance.

Salaried — asked once“Why we ask” travels with the question, not a help link
Irregular income, declared plainlyVerify with Singpass for a decision today — or upload a document and route to review
  • One question per screen — "Used only for this application. Never shown in your profile."
  • Ask with context, not suspicion: "Income forms part of the assessment. Your eligible Arta assets may also support your available limit."
  • Singpass, when the member chooses it: retrieve the IRAS Notice of Assessment via Myinfo — verified in ~30 seconds, decision today. The consent sheet shows exactly which fields, before authorising.
  • Adaptive friction, member-controlled: the same person can pick verify-now or upload-and-review — the evidence path changes, the rules never do.
  • No mid-form rejection: anything unresolved routes to review, never to a dead end.
09 / 19
Flow 1 · The application

Make the security and settlement mechanics explicit

Pay-in-full monthly · the security arrangement · what happens if values fall — always expanded, structure compliance-locked. Hard-pull consent in plain language. Exact eligibility, rates and security terms sit with Credit, Risk and Legal — the prototype says so on screen.

Eligible assets & recommended limitAssets stay invested; assessment logic on demand, marked illustrative
Settlement — a designed choiceThe charge card’s core promise, not a summary row
Key Terms never foldConsents separate and unchecked; full documents in Arta’s legal container
10 / 19
Flow 1 · The application

The decision is a relationship moment

Five governed states — approved · phased limit · more info · additional review · declined. The persona changes the evidence path, never the rules — and a declined applicant is still a wealth member tomorrow.

ApprovedThe newest card in your collection — virtual card usable now
Additional reviewThe evidence path differs, the rules don’t — portfolio already verified
DeclinedPrivate, specific as legally possible, zero cross-sell
11 / 19
Flow 1 · The application — The decisions behind it

The decisions behind Flow 1

Decision 01 · The income question

Keep one income question — even though assets secure the card

Weighed: skipping income entirely (the portfolio secures the card) vs. asking once.

Why this: a charge card still requires an affordability assessment — a regulatory obligation, not a design preference.

What would change my mind: Credit & Legal confirming that asset-based assessment alone satisfies responsible-lending rules.

Decision 02 · Prefill

Confirm, don't auto-submit

Weighed: auto-filling and submitting on the member’s behalf vs. showing verified data as review cards.

Why this: consent and data accuracy must be actively owned by the member — this is the one friction worth keeping.

What would change my mind: evidence that confirmation reads as busywork and hurts completion — then collapse it to a single summary confirm.

Decision 03 · Flow architecture

One flow with branches, not two journeys

Weighed: separate salaried and founder journeys vs. one flow that branches only on evidence.

Why this: the rules stay identical for everyone; only the evidence path changes — fairness stays legible.

What would change my mind: branching complexity leaking into the happy path: extra states, slower time-to-decision.

Decision 04 · Pledge scope

Member-selected pledge, not default-all

Weighed: pledging all eligible assets by default (fewer steps, bigger limit) vs. explicit member selection.

Why this: portfolio use must be explicit and member-controlled — the HMW tension, resolved toward control.

What would change my mind: members consistently pledging everything anyway, making the step read as ceremony.

How I'd validate — not yet run

Within this exercise I couldn't run research, so I designed the validation plan instead: completion & time-to-decision by evidence path · share choosing verify-now vs routed review · post-decline trust (support contacts + 30-day wealth retention after a "no").

Riskiest assumption, tested first: do members actually understand what securing assets means? A comprehension test on the pledge step comes before anything else.

12 / 19
Flow 2 · AI across the lifecycle — Principle 2 · Explain before acting

One AI relationship, not five features

The card gives Arta AI the one signal it never had — cash flow. Arta saw what you own; now it sees how money moves. Every touchpoint extends an agent Arta already has, and one moment is deliberately chosen for depth.

01 · Apply

Pre-qualified from data Arta already holds — and you can ask about any term, inline.

Confidence before committing
02 · Onboard

Autopay and statement date recommended from the member's real cash-flow patterns.

Set up right on day one, not defaults-by-inertia
03 · Everyday use

Spend insights in wealth context — burn vs. income vs. portfolio — with anomalies flagged.

"I understand my money"
04 · Manage & service

Repayment Intelligence — settlement sequencing, limit guidance, dispute triage with pre-filled evidence.

THE DEEP MOMENT · NEXT SLIDE
05 · Grow & retain

Rewards sweep into investments; an annual card-plus-wealth review, with real math.

The card compounds into the portfolio
13 / 19
Flow 2 · AI across the lifecycle

Deep moment: repayment intelligence — three realities, one anatomy

"What if markets move?" reprices her limit live — headroom transparency only a portfolio-secured card needs; "Why am I seeing this?" lists sources, calls itself arithmetic, and hands over the off switch.

QuietA normal month. The hardest call was letting the AI stay silent
NudgeStatement date moved to meet the founder’s invoice cycle
InterveneSettlement sequencing with a zero-cost fix — the capital-call line no bank can write
14 / 19
Platform · White-label — Principle 3 · Keep trust in the core

The core embeds. The shell is the partner's.

The line is enforced by the token pipeline, not by review: a partner theme that re-points a decision color to its brand pink fails the build. One chain: #a8cdf0 → ref.color.sky.40 → sys.action.primary → cmp.btn.primary.bg — Arta is simply theme #1.

The partner owns the relationship and the surrounding product shell; Arta protects the integrity of consequential financial decisions.

Partner owns

Relationship & shell

App navigation & entry point · brand identity, typography & color · card artwork · product naming · benefit presentation · tone of voice · support entry point.

Adjustable · guardrailed

Styled, within floors

Component styling · illustration · AI surface name · notification styling · explanation wording · information density · application entry point — each with a minimum the build enforces.

Arta core · non-removable

Consequential decisions

Eligibility & decisioning · eligible-asset logic & limit recommendation · required data model · compliance & disclosure hierarchy · consent & audit trail · security & authentication · AI safety boundaries · critical application states · card servicing.

Partner app shell  →  Partner-configurable experience  →  Arta-controlled card capabilities  →  Decisioning, compliance & servicing

LIVE STORYBOOK — 3 THEMES, GOVERNANCE STORIES ↗
15 / 19
Platform · White-label

Same bones, partner skin

Two fictional partners (Meridian, Northstar) prove the pipeline inside a partner-owned shell: the Arta core modules are identical everywhere — everything around them belongs to the partner.

ArtaKey Terms & consents
Meridian SecuritiesIdentical DOM — sys-token overrides only
Meridian — approval momentWithin motion & compliance guardrails
SIDE-BY-SIDE GOVERNANCE PROTOTYPE ↗
16 / 19
Closing

What's limiting in Arta's design system

And because a critique should be constructive: the missing layer now exists — 151 ref / 117 sys / 123 cmp tokens, the lending components, and three themes, running in a validated pipeline.

Gradients ≠ forms

Aurora surfaces fight data entry. Reserve them for moments — approval, card art — and add a quiet form surface.

No legal type style

Letterspaced caps micro-labels fail for disclosures. Lending needs a readable legal/tabular style.

Missing primitives

Dense inputs, chapter progress, currency input, consent rows, status timeline, document upload, decision states.

DESIGN-SYSTEM CONTRIBUTION ↗
17 / 19
Closing

If I had more time

  • Usability pass on the income & consent steps — comprehension testing on disclosure copy.
  • Card management & servicing surface — statements, disputes, freeze; same DS additions carry over.
  • Partner theming console — how a partner actually edits tokens with compliance-locked regions.
18 / 19
Closing

Everything in this deck is live.

Not mockups — the prototype, the partner view and the system are all running. Click through anything we discussed.

PROTOTYPE — FLOWS 1 & 2 ↗ WHITE-LABEL · NORTHSTAR ↗ STORYBOOK — 3 THEMES ↗ DESIGN SYSTEM DOCS ↗

Thank you.
Po Chun Lin · pochun.lin@thepacinglabs.com · linpochun.com

19 / 19