Case Study

AIA MOST BI Analytic Dashboard

A BI analytics tool that collapses 1,300+ Excel and PowerPoint trackers into a single, real-time management reporting surface for the AIA group finance team — shipped as an MVP in two sprints.

InsuranceBI / ReportingDesktop
My role
Design strategy · Product design (UX/UI)
Team
6 core · BA · Power BI specialist · UX/UI strategist
Outcome
4 dashboards · 8 weeks build · 30s → 3s render
Project overview

1,300+ trackers, zero shared view

The client — AIA's group finance team — wanted a data-driven, transformational BI solution to advance tracking across systems and projects. More than 1,300 systems and situations were tracked through Excel and PowerPoint, with constant back-and-forth to distribute, collect and clarify across reporting lines.

Driven by the business vision of MOST · Management One Source of Truth, the management consulting team brought EY Wavespace service design into the preliminary scoping conversation.

Challenge

“Analytica” was slow, fragmented and impossible to scale

A legacy self-service reporting capability that lacked any consolidated cross-market view, suffered slow renders, and wasn't portable to the Cloud — out of step with the group's wider cloud strategy.

No consolidated view

A self-service reporting capability for corporates that relied on data from multiple APAC countries — yet lacked any consolidated cross-market view.

30-second renders

Reports took up to 30 seconds to render, producing poor UX and weak adoption among the corporate users it was meant to serve.

Not cloud-portable

The legacy solution wasn't scalable and wasn't portable to the Cloud — out of step with the group's wider cloud strategy.

Approach

Design thinking folded into agile delivery

We leveraged design-thinking phases inside an agile delivery cadence — a four-week sprint pattern to take any new dashboard from Define / Prioritise → Co-design → Build → Test & Deploy.

Strategic-dashboard delivery process — design-thinking phases mapped into agile sprints.
Strategic-dashboard delivery process — design-thinking phases mapped into agile sprints.
Four-week agile delivery to an MVP dashboard — define / prioritise, co-design, build, test & deploy.
Four-week agile delivery to an MVP dashboard — define / prioritise, co-design, build, test & deploy.
Prioritisation & architecture

Start with the “why”, not the chart type

We deep-dived into the why — making sure the MVP was driven by user needs and motivations rather than aesthetics or trends. Exploratory workshops and stakeholder alignment surfaced the shared management-reporting vision across functions.

Exploratory workshops — best-practice scans across the insurance & BI industry.
Exploratory workshops — best-practice scans across the insurance & BI industry.
Stakeholder alignment — mapping management-reporting challenges into a shared BI vision.
Stakeholder alignment — mapping management-reporting challenges into a shared BI vision.
Define

Every data point traced to a calculation, a purpose and a place

Worked closely with data analysts to map out every data point and its calculation. Through categorisation we defined the purpose, data points and layout structure of all requested dashboards.

Data-point mapping — categorised by purpose and dashboard surface.
Data-point mapping — categorised by purpose and dashboard surface.
Layout structure — defined per dashboard ahead of co-design.
Layout structure — defined per dashboard ahead of co-design.
Co-design

A reusable BI design system, not just four pretty screens

Co-designed data dimensions, filters and page-level layouts. Generated templates and comparison patterns to be applied across future dashboard development — anchored by a BI style guide aligned to Power BI feasibility and the client's brand.

Visual style guide — colour, typography, composition & client-brand alignment within Power BI constraints.
Visual style guide — colour, typography, composition & client-brand alignment within Power BI constraints.
BI visualisation component library — reusable chart, filter and layout patterns.
BI visualisation component library — reusable chart, filter and layout patterns.
Low-fidelity wireframes — agreed with business on user flow across three levels of dashboard access.
Low-fidelity wireframes — agreed with business on user flow across three levels of dashboard access.
High-fidelity wireframes — design handoff for the Power BI engineering build.
High-fidelity wireframes — design handoff for the Power BI engineering build.
Click-through prototype — used to demonstrate the future experience to executive stakeholders.
Click-through prototype — used to demonstrate the future experience to executive stakeholders.
Outcome

From 2 years to 8 weeks. From 30 seconds to 3

The MVP collapsed render time by 10×, shipped in 8 weeks, and consolidated reporting into a single, cloud-ready source of truth.

30s → 3s
Render time

Report rendering dropped 10× — turning a coffee break into a glance.

8 weeks
To production

New solution built in 8 weeks vs. the original Analytica's 2-year rollout.

1
Single source

A single Operational Data Store significantly reduced support and maintenance cost.

Cloud-ready
Portable

Target architecture portable to the Public Cloud, in line with the group's cloud strategy.