Case Studies | Hospitality

Making public transport easier to navigate

AI Assistant

About

A Hospitality business working with Cloud Combinator on AWS. The client is anonymised at their request.

Challenge

The engagement concentrated on three focus areas that any AI enhancement had to satisfy in a safety-conscious transport context.

Reducing the cognitive load of the client's planning

Traditional planners tend to provide more information rather than better information, overwhelming users with timetables and maps. Over 16.35 million people over 65 and 4.5 million people with hidden disabilities in the UK can find these tools difficult, so the solution needed to turn complex, multi-modal journeys into clear, personalised guidance.

Keeping AI answers accurate and trustworthy

Applying generative AI to real-time transport is unforgiving: a hallucinated departure or connection erodes passenger trust. The design had to

Building on proven infrastructure, not from scratch

The client already integrates real-time feeds and popular messaging channels and processes behavioural data within an approved privacy framework. The AI layer had to slot into that mature platform with minimal new privacy infrastructure, reducing risk and speeding deployment.

Solution

The solution was structured as a phased development, carried out in partnership with the AWS GenAI team, so that AI capability could be proven and refined against real-world usage before wider rollout.

£348m

Value of currently untaken rail journeys the solution aims to unlock through better accessibility (projected)

7+ yrs

Of real the client's conversations used to fine-tune the transport models

Messaging channels already served: WhatsApp, Messenger and SMS

By the numbers:

£348m - Value of currently untaken rail journeys the solution aims to unlock through better accessibility (projected)
7+ yrs - Of real the client's conversations used to fine-tune the transport models
3 - Messaging channels already served: WhatsApp, Messenger and SMS

Changes

The designed solution combines the client's mature conversational infrastructure with generative AI to deliver personalised, accessible the client's guidance while keeping every journey-critical fact anchored to verified real-time data. The following outcomes are the targets the solution is built to achieve, developed in partnership with the AWS GenAI team.

Personalised the client's narratives — The solution transforms complex, multi-modal the client's data into plain-language, context-aware guidance tailored to each traveller.
Hallucination-resistant by design — AI interprets only verified real-time data from authoritative sources such as the platform and a national healthcare system, with factual content kept separate from AI explanation and outputs continuously validated.
Delivered on familiar channels — Guidance reaches travellers through WhatsApp, Messenger and SMS, building on infrastructure already serving thousands of daily users.
Designed for accessibility — The experience targets older travellers and people with hidden disabilities, reducing the cognitive barriers that keep many from using public transport confidently.
Built with AWS — The AI layer was developed in partnership with the AWS GenAI team, which committed development support and service credits to the engagement.

With the architecture designed and the phased plan agreed, the client is positioned to prove the AI layer with a closed test group and then scale through its operator relationships, using its GTFS-based platform as a route to wider UK and international deployment.