WEBJET · 2026 (Pending OpenAI Approval)

A native ChatGPT app for the next generation of travel search

We set out to be the first Australian Online Travel Agency with a native ChatGPT app.

UX Strategy, Product Design Direction

Webjet hotel results inside ChatGPT on a phone, with the prompt "@webjet Show me hotels in Bali"

The Problem

As user behaviour shifts toward AI-led searching, OpenAI's store has become a new starting point for travel research and we wanted to be there for it.

The brief had two halves, be early, and be useful. Leading design direction across hotels and flights and focused on a single principle: show just enough to build trust, surface the right options, and move the customer into Webjet's funnel ready to convert.

Customer Behavioural Shift

Customer behaviour was shifting faster than the industry was responding. Travel research was increasingly starting inside AI assistants customers letting these tools do the heavy lifting of comparing, summarising, and narrowing down before a booking site ever entered the picture.

For an OTA, that meant the moment of consideration was now happening on a surface we didn't own, with competitors we couldn't see. Being absent at that stage didn't just lose a click it lost the chance to be in the consideration set at all.

The Webjet app in ChatGPT on two phones, loading and then showing hotels in Bali
The same hotel flow in dark mode, loading and then showing rooms for a hotel
The loading state of the Webjet app in ChatGPT on desktop and phone
Hotel results from Webjet inside ChatGPT on desktop

The Tension

For this project, it wasn't how do we build for ChatGPT, it was understanding where the new customer's decision actually being made now.

That shift in framing changed everything that followed.

The Craft

With the moment of consideration happening on a surface we didn't own, the design had to do the heavy lifting immediately.

We designed the experience to surface high-value, relevant travel content upfront. By giving the user exactly what they asked for without the friction layer of traditional UI controls, we lowered the barrier to entry. Pulling them out of the research phase and into our conversion funnel with higher confidence.

Flight results from Webjet inside ChatGPT on desktop, with a link through to all flights on Webjet
A row of Webjet hotel cards with ratings, inclusions and prices
A row of Webjet flight cards with times, airlines and prices

The Solution

The work was as much about restraint as it was about design. No filters, no deep product detail, no second-screen experiences just clear, scannable answers that respected how people actually use these tools, with a deliberate path back to the full booking experience when the customer was ready to commit. It's early days for the surface, but the strategic logic is clear: as AI-led search grows, being present, useful, and trusted at the start of the journey is no longer optional it's where decisions are now forming.

Outcome

Currently pending OpenAI approval.

The work continues.
More shipped outcomes.

Different problem, same
uncompromising standard.

Next Project