Online travel: AI and digital delivery inside a live consumer business

What the work involved

AI and digital delivery in a consumer business that is always on. The work had to fit a high-traffic operation, which meant building for the constraints of a live system from the start.

Built for the constraints of a live system from the start, not a proof of concept sitting to one side.

Models and tooling went into the flow of how the business already runs, rather than sitting to one side as a proof of concept. That ran in two directions at once: shipping working software into production, and bringing the people who run the business along with it.

Bringing AI into the way the group works

The most durable part of the work was not a single piece of software. It was getting leaders across three separate companies in the group into the same conversation about AI, then turning that into a transformation model that could move quickly rather than stall in committee.

Start where the friction is, ship something real, build the appetite for the next step.

The first step was deliberately unglamorous: automating internal processes, beginning with HR workflows that ate time without adding judgement. That gave the organisation an early, visible win and a reason to keep going.

Underneath it sat practical, self-hostable tooling rather than a heavy platform purchase:

• Workflow automation to stitch together the repetitive, multi-system tasks the business was doing by hand.
• A self-hosted LLM chat workspace so teams across the group could use models safely, on the company’s own terms, without leaking data to consumer tools.
• AI-native development tooling rolled out to the engineering teams, with hands-on onboarding so the change stuck rather than being announced and forgotten.

The pattern throughout: meet teams in the tools they already use, prove value on a real problem, and let momentum, not a mandate, carry adoption.

A new consumer booking platform prototype

Built the customer-facing side of a new booking platform for a travel vertical (cruise): the full discovery-to-booking experience travellers actually touch. The work was deliberately scoped to the customer end ahead of the fulfilment side, getting a usable, revenue-capable front of house live first.

Strategy for an always-on travel assistant

Beyond delivery, we authored the product proposal for the group’s next-generation travel assistant: an always-on, context-aware companion that follows the traveller across channels rather than living inside a single app session.

The unlock: personalised travel help without giving up transaction simplicity.

The proposal set out the product vision, a high-level architecture for handling user actions and external “context events” (flight and price changes, disruptions) through a central agent, and a clear-eyed read of the competitive field. It framed several routes to value rather than betting on one:

• Direct to travellers: the existing audience, as the fastest path to commercial value.
• An AI-run travel desk for smaller businesses: a low-cost, efficient alternative to a full in-house travel department.
• A data-partnership position: protecting access to travellers even in a world where general-purpose assistants serve more of their needs.

Alongside it, a focused entry strategy for corporate and SME business travel: market sizing, a phased build-and-launch roadmap, go-to-market and partnership models, unit economics, and a risk-and-mitigation view: the case for a credible new B2B line built on existing capabilities.