Melbourne startup Alfred Travel has launched an AI-powered trip planner that builds itineraries, integrates with online travel agents, and promises future crypto rewards. But can it compete with incumbents and heavyweight LLMs?
Travel itineraries have long been touted as the perfect task for agentic AI. Alfred Travel, founded by Billy Ka Keung Chan and Andy Kwong Ching Tsang, is the latest to test the idea, pitching itself as a lightweight alternative to bloated booking engines.
Launched in July on iOS and Android, Alfred generates trip plans via prompts that are fed across multiple large language models, including OpenAI’s ChatGPT, Google Gemini, and xAI’s Grok.
It’s an API frenzy, generating prompts from simple user inputs (where and when you’ll fly, your budget, interests) and validating them with a string of LLMs before connecting through affiliate integrations with the likes of Trip.com, Traveloka, and Expedia. Restaurant recommendations are there but the booking systems integration are still in development, as is the blockchain pitch.
The plan is to issue “Alfred Tokens” to users who give feedback and engage with the app. Tokens could later be used for discounts or traded on exchanges. The tokens will also be used to unlock advanced features of the app.
Ka Keung, who previously worked in Hong Kong’s crypto sector at OSL, said the product came from seeing the potential in disrupting an industry that adopts tech slowly.
“Ten years ago we thought a one-stop shop for travellers was possible but the tech wasn’t there. Now AI can do the heavy lifting,” he said.
The Alfred Travel still feels raw: itineraries can be overly packed, there’s a notable lack of pricing transparency, a few odd bugs, and it’s not quite able to gather and act on the full context of a user prompt.
For example, Startup Daily tried to plan an itinerary for a trip to Melbourne for the AFL Grand Final weekend. While Alfred suggested ‘Sightseeing at Melbourne Cricket Ground’ on the Friday afternoon, it didn’t clock the Grand Final parade or try to suggest ways of getting a ticket for the last Saturday in September in the way ChatGPT did.
Of course, ChatGPT is an all-in-one chatbot made by a company valued at US$500 billion, whereas Alfred is three guys using AI to carve out a niche. And they’ve only been on the app store for a month.
According to Ka Keung, AI is everywhere from the data parsing through to marketing materials. It means the small, bootstrapped team can ship a reasonably well polished app and keep iterating as they prepare to scale. The startup’s “secret sauce”, he said, is its 576-line prompt, refined to produce the structured itineraries in one shot.
“The amount of work and the amount of cost that AI has saved us is incredible,” the Alfred Travel co-founder told Startup Daily. “Our runway is crazy. The month-by-month costline is 3 figures, under $1,000 per month. We can go for a very long time.”
Instead of raising venture capital, the Alfred team is banking on scaling its lean operation. Ka Keung said he’s hoping to hit 5,000 itineraries generated by the end of 2025, with breakeven forecast at around 15,000 itineraries, generously assuming 10% of those converts to bookings.
“We’re very conscious about cost because we want to own most of the business, but we are interested in strategic partnerships,” Ka Keung said. “If Flight Centre said ‘hey we’ll give you a buck for a stake’ maybe we’d be interested.
“The way we see it, why couldn’t Flight Centre agents themselves use Alfred? Instead of digging through databases, they could print out an AI-generated plan and focus on the customer relationship.”
Daily startup news and insights, delivered to your inbox.