Business travel for the AI Age
We took a 20-year-old publicly traded hospitality brand into the AI Age, and the American market. What started as a 6-week engagement, turned into the LCA team defining, designing and building a whole new agentic product.
From vision to full agentic product build
A frontier embedded team from vision to production.
As a deeply integrated product partner, we took Serko, a legacy technology and travel brand, to Serko AI in a matter of months.
The LCA team ran strategy, brand, product design, and engineering together, from the first vision sprint through to the shipped code.
I want a flight leaving at 3:45pm
I need to be there by dinner
Understand “why” you’re traveling, and infer what you need from that.
Going to Chicago for GBTA, arrive Sunday evening, leave Wednesday afternoon
Building an itinerary based on your preferences…
- Hotel only (no apartments)
- Prefer closer to destination
- Marriott brand
- Hilton brand
- Wi-Fi room
Flights are nonstop-heavy and line up well with your American Airlines preference. Hotels stay within your company’s nightly guideline.
Want to refine anything before you choose?
Tailors every choice to you and your company’s preferences.

Takes action for you, clearly and confidently.
An agent is only as good as the context it can get.
Our strategy work turned insights into the foundation of the Serko AI experience. Every trip has multiple reasons behind it, and the agent uses those as context to make the recommendations that fit a traveller's goals.
The agentic booking experience.
Describe your trip with a sentence. Where you're going, and why it matters. Serko AI asks for any other needs, then makes searches based on your intent, personal preferences, and company policy. Your trip is assembled as you talk, so you never lose the whole picture inside a long thread.
You approve, it books, and remembers what you chose for next time. Each trip teaches it a little more about your needs, and earns a little more trust to act on its own.
We engineered the agent to show its work.
Behind the conversation, Serko AI orchestrates an LLM context agent, policy extraction, and parallel calls to supplier APIs, then merges and scores the results that come back.
Each step was built to be observable. Every conversation is captured as a trace, tool calls and agent handoffs are checked against expected behavior, and regression evals run before changes ship. Allowing the team to pinpoint how an agent made a decision, and why.
