12am late night in UTS Claire, Magic and Veralyn after the pitch

Capgemini × UTS Consulting Club

My Role : UX Strategist

My Team : UX Researcher, Strategy Analyst

Awards 🏅:

UTS x Capgemini 2026 - Semi Finalist

PROJECT OVERVIEW

Interstellar Bank: Digital Transformation Strategy

A five-year strategy to reduce loan processing friction and improve digital adoption.

We developed a transformation roadmap for Interstellar Bank's loan processing division, using competitor benchmarking, operational diagnosis, cost modelling, ROI projection, and risk planning to propose a unified loan origination system: OneFlow.

15 days → target under 5 days A$615K estimated investment ~2-year break-even 139% projected ROI
Manual verification
Fragmented workflow
Outdated loan platform

Proposed system

OneFlow

Faster approvals
Operational visibility
Staff adoption
BUSINESS CONTEXT

The bank had a growth problem hiding inside its operations

Interstellar had lending potential, but slow internal processes limited its competitiveness.

Interstellar Bank had a strong home lending base, but its loan processing division was slowed by manual verification, fragmented workflows, outdated systems, and post-acquisition misalignment. The business problem was not simply a lack of digital tools — it was an operating model that could not scale efficiently.

3.5M

Existing customers across retail and business banking

A$80B+

Home loan portfolio — the bank's core revenue base

~15 days

Average loan processing time — 3× slower than competitors

1 acquisition

Lunar Savings, adding system and cultural complexity

Why this mattered

The acquisition wasn't just a growth play — it created operational debt. Two different systems, two different cultures, and a management team still figuring out how to run one bank. That's where the real problem was hiding.

Key insight: Interstellar did not just need a new digital interface. It needed a more connected operating model across process, technology, and people.
PROBLEM DIAGNOSIS

Loan delays were a symptom — the operating model was the problem

Our diagnosis showed that loan delays were caused by more than manual admin. Interstellar's process was affected by fragmented workflows, system crashes, outdated platforms, data sync errors, unclear leadership, staff skill gaps, and inconsistent adoption across teams.

01

Process

Manual verification, slow approvals, duplicated work, and unclear handoffs across the loan journey.

02

Technology

Outdated loan platforms, data sync errors, unreliable systems, and limited workflow visibility.

03

People

Post-acquisition misalignment, staff skill gaps, low morale, and resistance to new digital workflows.

Key diagnosis: The problem was not "lack of an app". It was a disconnected operating model across process, technology, and people.
COMPETITOR BENCHMARKING

Benchmarking showed where Interstellar was falling behind

Competitors were faster, and speed was tied to customer experience.

We benchmarked Interstellar against major Australian banks and found a clear performance gap. Interstellar's estimated loan processing time was around 15 days, compared with approximately 5 days for leading competitors. This suggested that loan speed was not just an internal efficiency metric — it affected satisfaction, trust, and competitiveness.

Estimated loan processing time

Days
5
10
15
5.5 days

CommBank

5.0 days

Westpac

15.0 days

Interstellar

i
Manual document handling and fragmented legacy systems contributed to processing times estimated at 3× the industry average — the key driver behind the OneFlow proposal.

Estimated figures based on benchmark analysis from the project deck.

Interstellar was approximately 3× slower than leading competitors, making loan processing speed a strategic priority rather than a back-office issue.
So what

If Interstellar wanted to compete in home lending, improving loan processing speed had to become a business priority — not just an operational clean-up task.

STRATEGIC OPPORTUNITY

The opportunity: make loan processing a competitive advantage

The strategy needed to improve speed, visibility, adoption, and operational control.

The research pointed to one clear direction: modernise the loan origination process while helping staff adopt the new workflow. We framed the opportunity around two connected layers — a unified system for processing loans and a people strategy to support behavioural change.

How might we

How might Interstellar reduce loan processing friction while helping staff and customers confidently adopt a more digital lending experience?

Operational friction
Strategic response
Click any row to see root cause and expected business impact →
01Manual verification
AI-assisted document extraction and verification
Root cause: Loan officers manually reading and re-entering data from uploaded documents — taking 2–3 days per application.
Expected impact: Reduce document processing from days to under 2 hours with AI extraction at ~94% accuracy.
02Fragmented workflow
Unified loan dashboard
Root cause: Staff using 4+ disconnected systems — email, spreadsheets, legacy LOS, and acquired bank tools — with no single view of loan status.
Expected impact: Single source of truth reduces handoff delays, cross-team miscommunication, and duplicate data entry.
03No live status visibility
Real-time approval tracking
Root cause: Customers had no way to track their application after submission — generating high call centre volume and trust erosion.
Expected impact: Self-serve status tracking reduces inbound enquiries and improves perceived transparency.
04Staff adoption gap
Training, onboarding, and change champions
Root cause: Post-acquisition staff from Lunar Savings using different workflows. No shared training or incentive to adopt new systems.
Expected impact: Phased training with internal champions reduces resistance and accelerates platform adoption.
05Missed cross-sell opportunities
Contextual product recommendations at approval touchpoints
Root cause: Approval was a terminal event — no system-level prompt to surface relevant products at the point of highest engagement.
Expected impact: AI-driven cross-sell recommendations at approval touchpoints increase revenue per loan application.
Closing insight: The solution could not just automate tasks. It had to connect people, systems, and decision-making into one operating flow.
THE PROPOSED SYSTEM

OneFlow: a unified loan origination system

A connected operating system for customers, staff, and loan operations.

OneFlow was proposed as a unified loan origination system that connects customer intake, document verification, credit assessment, workflow prioritisation, approval tracking, and staff dashboards. The goal was to reduce manual handling while creating clearer visibility across the loan journey.

Step 01

Customer submits loan application

Step 02

OneFlow extracts and verifies documents

Step 03

Risk and urgency are prioritised

Step 04

Staff review exceptions

Step 05

Customer tracks approval status

Step 06

Feedback loop improves operations

System layer

AI-assisted document extraction
Centralised loan dashboard
Real-time status tracking
Workflow prioritisation
Exception handling

Adoption layer

Staff training
Change champions
Phased onboarding
Leadership alignment
Continuous feedback loop
OneFlow was designed to improve the operating flow, not just digitise the existing process.
WORKFLOW PROTOTYPES

Making the strategy tangible through workflow prototypes

The prototype showed how OneFlow could change both the customer and staff journey.

To make the transformation strategy easier to understand, we created customer-facing and staff-facing prototype touchpoints. The customer flow showed how digital onboarding, document upload, credit assessment, real-time approval tracking, and AI support could reduce uncertainty. The staff-facing dashboard showed how teams could monitor platform status, prioritise applications by risk and urgency, check document completeness, and identify opportunities at the point of approval.

What we tested

We weren't testing whether the UI looked polished. We were testing whether the workflow logic made sense to a stakeholder audience — could someone follow the loan journey end-to-end just by looking at these screens?

Group 1

Customer-facing flow

Reduces uncertainty during the loan application journey.

Let's get started

Personal loan · about 10 minutes

Pre-filled from your banking profile — 8 fields auto-completed. Review and confirm.

Loan Details

Loan amount

$100,000

Loan purpose

Home renovation AI

Term

36 months

Repayment frequency

Monthly

Your Details

Full name

Sarah Mitchell

Continue →
Action required

Upload your documents

📄

Photo ID — Passport

Identity confirmed

Done
💼

Payslips (3 months)

Income $6,200/mo extracted

Done
🏦

Bank statements (90 days)

AI scanning transactions

Scanning
📋

Employment letter

Not yet uploaded

Upload
✦ AI extraction live ● Live
EmployerDeloitte Australia
Monthly income$6,200
Employment typePermanent full-time
Pay cycleFortnightly
ABN74 599 608 295
✓ 0 risk flags detected
AI assessment

Credit assessment

742 Equifax Score
Excellent

Top 15% of applicants

2.5

DTI Ratio

0

Missed Payments

4 yrs

Employment

✦ AI risk summary 94% confidence

Your profile outperforms 85% of applicants this month. Income stability over 4 years with zero gaps detected. First-time borrowing is your only flag, offset by strong employment tenure and a DTI of just 2.5.

In progress

Under review

Our team is finalising your application

REAL-TIME TRACKER · #INS-2024-08843

Documents verified

Today 10:42 AM

Credit check completed

Today 10:45 AM

Application under review

In progress · est. 2 hours

Compliance sign-off

Pending · ~1 hr

Final approval decision

Pending

↑ Moving 8 mins faster than average — same-day approval likely

Estimated Decision

Today by 3:00 PM

Based on current queue and your profile

AI Loan Assistant

Online
Talk to human

⚡ Powered by Interstellar AI

AI

Hi Sarah 👋 I'm your Interstellar loan assistant. Your application #INS-2024-08843 is in workflow review. How can I help?

How long will the review take?

AI

Based on your profile and today's queue, I estimate 1–2 hours. You're moving faster than average — documents were clean and credit cleared in 1.2 seconds. I'll push a notification the moment a decision is made.

Can you resend the employment letter link?

AI

Done! Sent to sarah.mitchell@email.com and a push notification to your device. Tap below to upload directly.

↑ Upload employment letter

Group 2

Staff-facing workflow

Gives teams visibility, prioritisation, and operational control.

Live
LO

Lisa Okafor

Sr. Loan Officer

Dashboard
Intake Queue 47
Doc Review 8
Credit Review 5
Flagged Cases 3
Approvals 6
Disbursements 4
AI Assistant

Good morning, Lisa · Friday 11 April 2025

Dashboard

Data Migration

90%

Complete w/o downtime

LOS Platform

Live

Across all teams

Staff Trained

80%

+12% this week

System Uptime

99.5%

0 incidents

Loan Approvals

<4 wks

Avg 2.3 days

Live Queue ✦ AI triaged View all →
ApplicantAmountActionSLA

Sarah Mitchell

#INS-2024-08843

$25,000

Personal

Review 4h 22m

James Kowalski

#INS-186842

$80,000

Business

Review 2h 48m

Aisha Lim

#INS-186943

$12,500

Personal

Urgent 1h 12m

Paulo Reyes

#INS-168842

$45,000

Vehicle

View Done
Live

Risk & Urgency Queue

Sorted by AI risk score · SLA priority

Aisha Lim

#INS-186943 · $12,500

HIGH RISK

Marcus Webb

#INS-186901 · $67,000

HIGH RISK

James Kowalski

#INS-186842 · $80,000

MEDIUM

Yuki Tanaka

#INS-186812 · $34,000

MEDIUM

Sarah Mitchell

#INS-08843 · $25,000

LOW RISK

Paulo Reyes

#INS-168842 · $45,000

LOW RISK
Live

SLA Countdown

Time remaining per application

Aisha Lim · #INS-1869431h 12m left
James Kowalski · #INS-1868422h 48m left
Sarah Mitchell · #INS-088434h 22m left
Marcus Webb · #INS-1869016h 05m left
Live

Document Completeness

Colour = verified / missing / pending

ApplicantIDIncomeBankEmploy.
Sarah Mitchell
James Kowalski
Aisha Lim
Paulo Reyes
Marcus Webb
Live

Exception Review

Flagged items requiring manual review

CREDIT

Aisha Lim · #INS-186943

DTI of 68% significantly exceeds 40% limit. 4 missed payments in past 12 months. Manual review required.

INCOME

Marcus Webb · #INS-186901

Income verification failed — employer ABN not found in ATO registry. Document re-upload requested.

ID

Yuki Tanaka · #INS-186812

Passport expires within 6 months. Secondary ID requested for verification.

ADDRESS

Sarah Mitchell · #INS-08843

Address on ID does not match bank statement. Minor discrepancy — flagged for staff confirmation.

This prototype was designed to communicate the proposed workflow and strategy, not to represent a production-ready banking platform. All screens show hypothetical user interactions based on the proposed OneFlow system.
FINANCIAL FEASIBILITY

Testing feasibility through cost, savings, and ROI

The recommendation included a high-level financial model, not just a feature proposal.

We estimated the implementation cost of OneFlow at A$615K, covering system development, integration infrastructure, compliance tooling, and training and change management. With projected annual operational savings of A$350K at full run-rate, the model estimated a break-even point of around two years and a five-year ROI of 139%.

A$615K

Estimated investment

Hover for details ↑

How we estimated it: Development (A$325K), integration (A$150K), compliance tools (A$85K), and training (A$55K) — modelled from comparable LOS implementations and project scope.

A$350K

Annual savings

Hover for details ↑

How we estimated it: Reduced manual processing labour (~A$180K), lower error rework costs (~A$90K), and reduced call centre volume (~A$80K) at full run-rate from Year 3.

~2 years

Break-even

Hover for details ↑

What this means: Cumulative cash flows turn positive between Year 2 and Year 3 based on phased investment and ramping savings. Modelled as directional — not a guaranteed financial outcome.

A$855K

Five-year net benefit

Hover for details ↑

Why it matters: This is the cumulative net position after five years — savings minus total investment. A$855K net benefit over five years demonstrates the financial viability of the transformation.

139%

Projected ROI

Hover for details ↑

How it's calculated: (Net benefit ÷ Total investment) × 100 = (A$855K ÷ A$615K) × 100 = 139%. This is a five-year return, not annualised.
Model assumptions
Investment horizon 5 years Phased across Years 1–2
Annual savings run-rate A$350K/year From Year 3 at full implementation
Year 1 savings (partial) A$175K 50% run-rate during pilot phase
Discount rate Not applied Figures are nominal, not discounted
Basis High-level estimate Not a formal audited financial model

Cumulative net cash flow

A$855K A$555K A$250K A$0 −A$220K −A$220K −A$55K A$250K A$555K A$855K Year 1 Year 2 Year 3 Year 4 Year 5 ≈ break-even

Investment breakdown

Development & automation A$325K
Integration & infrastructure A$150K
Compliance & risk tools A$85K
Training & change management A$55K

Figures are high-level estimates from the project model and should be read as directional, not audited financial forecasts.

IMPLEMENTATION ROADMAP

A strategy is only credible if it can be implemented

We mapped a phased rollout across technology, people, and operational change.

The five-year roadmap balanced system rollout with organisational change. The plan moved from system design and training, to pilot testing, full launch, optimisation, and AI scaling. In parallel, the people and culture stream focused on value alignment, leadership unity, team incentives, and building a more integrated post-acquisition culture.

Year 1
Year 2
Year 3
Year 4
Year 5
Technology

System design and architecture

Technology

Core build and pilot testing

Technology

Full launch

Technology

Optimisation and workflow refinement

Technology

AI scaling and continuous improvement

People

Staff training foundation

People

Value alignment

People

Leadership unity

People

Team incentives

People

One-team culture

The roadmap treated adoption as a core workstream, not an afterthought.
RISK CONSIDERATIONS

Designing for adoption, migration, and risk

The strategy considered what could fail before rollout.

A recommendation is only as strong as its failure modes. Before presenting the strategy, we stress-tested it against three categories of risk: how staff might resist it, what could go wrong during data migration, and where AI-assisted verification could fall short. Each risk shaped the design of the rollout rather than being listed as a caveat at the end.

Adoption risk

Staff resistance

Post-acquisition cultural tension between legacy bank staff and the new parent organisation could slow uptake and undermine the unified platform's effectiveness.

Mitigation

Phase in change champions early, tie incentive structures to platform usage milestones, and use Year 2 of the roadmap to focus on value alignment rather than system features.

Migration risk

Data sync and migration errors

Moving loan records, customer data, and credit histories across systems creates a window for corruption, duplication, or loss — especially given the acquired bank's fragmented legacy infrastructure.

Mitigation

Run parallel systems during the Year 2 pilot period. Require data reconciliation audits before sunsetting any legacy system. Scope rollout branch-by-branch to limit blast radius.

Verification risk

AI-assisted verification limits

Automated document checking and credit assessment could miss edge cases or introduce bias — creating compliance exposure and undermining customer trust if decisions feel opaque.

Mitigation

Human review gates at high-risk decision points. Regular model audits for bias and drift. Staff dashboard designed to make AI reasoning visible rather than hiding it behind a score.

Risk planning shaped the roadmap sequence — it wasn't added after the strategy was set.
OUTCOME

Outcome: a transformation strategy with measurable business logic

The final recommendation connected operational pain points to cost, rollout, risk, and adoption.

The final output was a five-year transformation roadmap for Interstellar Bank's loan processing division. The recommendation addressed manual processing, fragmented workflows, digital adoption challenges, and customer uncertainty through a proposed unified loan origination system and change enablement plan. The business case projected a two-year break-even period and positive five-year return, while the rollout plan accounted for implementation risk and organisational adoption.

01

Strategic diagnosis

Identified bottlenecks across process, technology, and people.

02

Competitive benchmark

Compared Interstellar's loan processing performance against leading banks.

03

Business case

Modelled estimated cost, savings, break-even, and ROI.

04

Implementation plan

Proposed a phased roadmap with adoption and risk controls.

A$615K Estimated investment
~2 yrs Projected break-even
A$855K Projected five-year net benefit
139% Projected ROI

Projected outcomes were based on the project model and were not live implementation results.

REFLECTIONS

Strategy is only as credible as the thinking behind it

12am late night in UTS
12am late night in UTS
Claire (left), Magic (right)
Claire (left), Magic (right)

What I learned

Working through the financial model taught me to question every number before committing to it — a habit that now shapes how I approach any design decision.

What was challenging

Digging deep into research to justify every number we presented. Unlike typical design work where you're solving for users, here we were building a recommendation for industry stakeholders — which required a completely different way of thinking and communicating.

What I'd do differently

Invest more time upfront in stakeholder mapping. We focused heavily on the operational diagnosis, but discovered late that understanding who needs to be convinced shapes every framing and design decision along the way.