Connie Giannopoulos
Selected work
UX/UI·AI·Research

Ask The Australian

AI-powered search built for business journalism.

Ask The Australian
CompanyNews Corp Australia
RoleSenior Product Designer
TimelineAugust - October 2025
LaunchApril 2026

Overview

Ask The Australian is a generative AI search experience built for the Business section of The Australian. It combines AI-generated answers with sourced journalism, giving subscribers a faster, smarter way to find insight without sifting through search results..

The challenge

The Australian has a highly engaged, opinionated readership with high expectations. One clear friction point was content discovery. Valuable journalism was getting buried and the existing search experience wasn't helping readers find it. Prior to the project, a survey of over 500 subscribers confirmed it: faster, smarter content discovery was the top personalisation priority.

This mattered most for two subscriber groups:

01
Nation Builders

Business leaders and entrepreneurs who need reliable information quickly. Skew younger: 24% Gen Z, 36% Millennials.

02
Older Affluents

Experienced investors after detailed, fact-based research without sifting through pages. Predominantly Gen X, aged 45–65.

The project came with clear constraints. The brand identity needed to be respected, legal guardrails ruled out financial advice and personally identifiable information, and the first iteration was scoped to web only. Anything technically out of reach was documented for future iterations.

Defining the experience

Early on the stakeholders made a call that this was a Gen AI search, not a chatbot experience. The goal was to help people find answers quickly, understand where those came from, and keep reading. That distinction shaped the information architecture, the interaction model, and the way content was presented.

Gen AI search, not a chatbot

A chatbot has a conversation. Gen AI search generates a single, direct answer from a trusted source and gets you on your way. No back and forth, no open ended dialogue.

Built on verified journalism

Rather than drawing from the open web, Ask The Australian pulls answers from The Australian's archive. Every response is sourced, cited and grounded in credible reporting.

Initial experience mapping

From there I mapped out the core journey, covering how users would discover the feature, what they would encounter, and where they could end up.

A five-column experience map showing the user journey from customer profiles through to outcomes including reading source articles, asking follow-up questions, exploring related prompts, and providing feedback.
Initial experience mapping · Customers, Discover, Start, Discovery, & Outcome

Early concepts

The first round of designs focused on the core journey within The Australian's established brand.

A grid of ten mobile screen designs showing the V1 Ask The Australian interface, including the welcome screen with personalised greeting, market data with graph, AI summary, sources and related prompts, search results, feedback flow, and all sources.
First round of mobile designs covering the core journey and key interaction states

Research & findings

With the first round of designs complete, I worked alongside a Senior UX Designer to run one on one user interviews with subscribers from both audiences. Before sessions began, a testing flow was mapped to define the exact paths participants would move through and structure the research around the right questions.

A flowchart showing the structured user testing flow with two entry points, prompt selection, AI summary interaction, thumbs down feedback path, negative response state, market results, and a step-by-step interview timeline at the bottom.
Structured testing flow with interview steps

One theme came through consistently: users preferred the AI feature over traditional keyword search. It saved time and surfaced content they wouldn't have found otherwise.

Key Findings

01
Trust & accuracy

Users trusted the AI because it drew from The Australian's journalism rather than the open web. But trust was fragile. Summaries that felt incomplete eroded confidence quickly.

02
Visual market data

Users wanted charts, graphs and customisable timeframes, not descriptive text. If it couldn't be built to that standard, it shouldn't be there.

03
Source transparency

Sources were valued, but users wanted attribution linked directly within the summary itself.

04
Related prompts

Users saw related prompts as a natural way to keep exploring.

05
Entry point discoverability

The banner didn't create enough urgency, and the floating icon alone didn't communicate the AI function.

06
Personalisation expectations

The greeting set expectations the product couldn't meet as it wasn't in scope, so it was removed.

Refined user flow

With the research synthesised, I updated the user flow to map every response state and exit point before moving into V2 designs.

detailed user flow diagram showing paths from The Australian Business Page through two entry points, into the AI search popup, across AI summary, partial response, and negative response screens, with branching paths for sources, related prompts, feedback, and article navigation.
Finalised user flow reflecting all response states and exit points

Iteration

Research informed a focused set of design decisions. Rather than changing everything, I focused on the moments that mattered most: how users discovered the feature, what they encountered when they opened it, and how the AI communicated its sources.

I also refined the visual styling throughout, simplifying the UI to give the feature a more polished and distinct feel within The Australian's brand.

V2 mobile screens showing the refined business landing page with a stronger banner and orange pill entry point, alongside the updated popup with dark teal header, refined search bar, and surfaced disclaimer.
v2 · landing page and pop-up
V2 mobile screens showing the partial response state with transparent AI limitations and related results, alongside the structured feedback flow with checkbox selections and optional written input.
v2 · partial response and feedback example

Source attribution came through strongly in research but couldn't be built within scope. The first screen shows the shipped V2 experience. The following two are future state explorations of what inline citation could look like.

Three mobile screens showing source attribution across V2 and two future state explorations. The first shows the shipped V2 AI summary with sources listed below. The second shows inline numbered citations linked to source cards with images. The third shows a selected citation expanding to reveal the full source, byline, and supporting quote.
v2 · ai summary and further explorations

Prototype

The prototype covers the end to end experience: launching the feature, exploring suggested prompts, running a search query and interacting with the feedback flow.

Ask The Australian · Figma make prototype

Reflection

Ask The Australian launched in April 2026 as News Corp Australia's first AI-powered search experience. Post-launch data wasn't available at time of writing.

What worked

Getting designs in front of real subscribers early and feeding those insights directly back into the product.

What I learned

Designing for trust isn't just about accuracy. For an audience this sceptical, every small decision either builds or erodes it.

What I'd do differently

I'd have pushed harder on source attribution earlier. It came through strongly in research, couldn't be built within scope, and is the clearest next step for the feature.

All designs shown are my own work and may differ from the live product.
NextRateCity Design System