< 2 weeks
To integrate OpenUI Cloud
To integrate OpenUI Cloud
Compared to A2UI
Frontend deployments
Working primarily with banks, telecom providers, and travel companies, Point Labs combines loyalty infrastructure, AI-powered engagement, and rewards marketplaces to help organizations drive customer retention, increase product usage, and create more personalized customer experiences.
One of those deployments involved building a voice-first AI travel concierge for one of Asia's leading banks. Deployed as part of the bank's loyalty program, the solution helps customers discover flights, hotels, offers, and loyalty benefits through natural conversation.
The goal was simple: make travel booking as easy as having a conversation. Customers could speak naturally to an AI assistant, asking questions like:
"Find me flights to Mumbai tonight."
The system would search flights, hotels, offers, and loyalty benefits, then recommend the best options based on the customer's request.
But while AI could generate recommendations, presenting those recommendations effectively was a different challenge.
Travel booking is inherently visual. Customers need to compare options, browse hotels, evaluate flights, and make decisions quickly. A wall of text was never going to provide the experience Point Labs wanted to create.
The team needed a way to dynamically generate interfaces that could adapt to every customer query while keeping the conversational experience intact.
Initially, Point Labs built the experience using A2UI. While it helped the team get started, they quickly ran into architectural limitations.
Slower responses. Every time a customer asked a question, the model had to process both the travel request and the instructions needed to generate the interface. As a result, generating responses became slower than the team wanted.
"The problem is that for each and every request to generate a UI, the entire UI schema definition goes to the LLM. The LLM is getting both the rules to do the reasoning of my business process and also generate the UI."
~ Shamsundar Pramanik, Head of Engineering
The team also faced workflow challenges. Making even small changes to the user experience required updates to the AI system itself.
Because UI definitions lived inside the AI backend, even small presentation changes required involvement from AI engineers and backend deployments, making routine UI updates more complicated than they needed to be.
To solve these challenges, Point Labs adopted OpenUI Cloud. Rather than asking a single system to handle both travel recommendations and interface generation, the team separated those responsibilities.
The AI focused on helping customers discover the best flights, hotels, and offers, while OpenUI Cloud handled how that information was presented.
For Point Labs, this meant faster responses, greater frontend flexibility, and a more maintainable development workflow.
The result was a travel booking experience that felt less like chatting with an assistant and more like browsing a purpose-built travel application, complete with interactive recommendations, booking options, and dynamically generated interfaces.
Moving to OpenUI Cloud improved both the customer experience and the way Point Labs built it.
67% lower latency in response times. By separating travel intelligence from interface generation, Point Labs achieved lower latency compared to its previous A2UI implementation. Faster responses made the travel concierge feel more natural and responsive, particularly when customers were exploring multiple destinations, flights, hotels, and offers in a single conversation.
0 dependency frontend deployments. With OpenUI Cloud, Point Labs no longer had to route UI changes through the AI system. This gave frontend teams greater ownership of the user experience and allowed AI engineers to stay focused on improving travel recommendations and booking workflows.
"I want to have a specific frontend developer. I want to have an AI engineer who is doing the backend stuff. In A2UI there is no segregation."
~ Shamsundar Pramanik, Head of Engineering
With frontend and AI development no longer tightly coupled, Point Labs gained the flexibility to continuously refine how travel recommendations were presented to customers. Instead of navigating booking options through static text responses, users could browse flights, compare hotels, explore offers, and interact with recommendations through dynamic interfaces tailored to each request.
For Point Labs, the goal was never simply to add AI to the travel booking experience. The goal was to make travel planning feel as natural as having a conversation.
Following the successful deployment in Sri Lanka, Point Labs has begun exploring similar rollouts across additional markets in the Middle East, Indonesia, and Europe.
By combining agentic AI with Generative UI, the team was able to move beyond both traditional booking flows and text-only chat experiences. Customers can now explore flights, hotels, rewards, and offers through interfaces that adapt dynamically to each request, making it easier to discover options and act on them.
"For an LLM on the other side to decide what is the best UI that can be put in front of the user — I think that is the core idea that you guys are solving."
~ Shamsundar Pramanik, Head of Engineering