AI-First Development

Laying the Foundations for AI-First Development

2
days hands-on AI workshop
3
core outcomes delivered
Weeks
to a live support agent

About the Client

Traveltek is an established travel technology provider, building booking solutions for travel agents, tour operators, and wholesalers. Like many businesses in their position, they wanted to explore how AI could speed up development and drive product innovation without disrupting the workflows that already worked well for them.

The Challenge

In early 2025, Traveltek set out to become an AI-first travel technology company. The ambition was clear, but they needed a practical way in. Their engineering and product teams needed to understand agentic development and how tools like Claude Code and Model Context Protocols (MCPs) could fit into their processes. They also needed to find the automation opportunities that mattered most to their core products, particularly within cruise and support, and they needed tangible results quickly to build confidence across the business.

How We Helped

We ran a two-day, hands-on AI workshop with Traveltek’s engineering and product leadership. Rather than following a formal framework or maturity assessment, the sessions centred on live demonstrations and collaborative exploration.

We showed how Claude Code could analyse large codebases, make small, reviewable code changes, and generate documentation and tests in a controlled way. The teams learned how to use structured goals, planning loops, and human-in-the-loop checks to build trust in AI-generated output. Working together, we also built rich, visual cruise quotes using AI, right there in the room.

From there, we moved on to what came next. Together we mapped out the architecture and workflow for Traveltek’s first AI-powered Support Investigation Agent, covering trigger points, scope, and guardrails. In the final session, we explored how AI could improve cruise discovery, an idea that became the Cruise AI Search concept.

What We Achieved Together

The workshop delivered three clear outcomes:

  • Foundations for the Support Investigation Agent, which Traveltek implemented shortly afterwards
  • Rich, visual cruise quotes, built live with the Traveltek team
  • The Cruise AI Search concept, which we later prototyped and handed over for Traveltek’s internal deployment

Since then, Traveltek’s teams have applied the same AI-first methods across internal tools, improving turnaround times and reducing repetitive support work. We’ve continued to collaborate on expanding the Support Agent’s capabilities, building the full Cruise AI Search product, and developing an internal AI Enablement Playbook to track maturity and return on investment across future pilots.

What This Means Going Forward

A few lessons stood out. Context comes before autonomy: the success of AI agents depends on investing time early in clear documentation and well-defined constraints. Tangible wins drive adoption, and seeing real results in the room created momentum and buy-in. And designing for handover matters, because packaging prototypes cleanly meant Traveltek’s engineers could keep building independently.

The partnership established a proven, repeatable model for applying AI safely and effectively in travel technology, helping Traveltek move faster while keeping confidence and control.

Team discussing ideas around a white table with laptops
Want results like these?
We want to hear more about your project.
Case Studies

Different clients, same
high-quality outputs