Islamic Lifestyle 

New simple chatbot start of more intelligent tech for Singapore’s Have Halal, Will Travel

| 06 June, 2018
 Yosi Winosa
New simple chatbot start of more intelligent tech for Singapore’s Have Halal, Will Travel
Photo: Have Halal, Will Travel co-founder Mikhail Melvin Goh and Rachel Tan, managing director of tech partner PleoData. Photo supplied by Have Halal, Will Travel.

The simple version of a chatbot on online travel guide Have Halal, Will Travel (HHWT) is the start of a more intuitive system that will in future be able to respond to users’ queries with deeper understanding, Mikhail Melvin Goh, HHWT co-founder told Salaam Gateway.

Late last month, HHWT, in partnership with tech firm PleoData, launched its chatbot Sofia—the name picked by most users out of a choice of six—using Microsoft Azure tools in combination with language understanding intelligence services (LUIS).

According to Goh, HHWT wanted to come up with a tool to match its users’ digital lifestyle. “The question was: how can we serve our 9 million users looking for halal food options over platforms like Facebook, e-mail or website comments every day?” Goh told Salaam Gateway.

 “The answer was obvious: we needed a chatbot. Then we created Sofia simply because people have been asking us questions which we found could be answered largely through automation,” he added.

The Singapore-headquartered company said its approximately 9 million users a month are primarily based in Singapore, Malaysia and Indonesia.

The Azure cloud platform comes with an application programming interface (API) for a question and answer maker, which was the first building block the HHWT team tapped into to form answers to commonly-asked user questions. Microsoft’s LUIS then enables chatbot designers to draw on datasets already in existence, enabling the team to analyse what users would ask.

Goh said the chatbot took HHWT a couple of months to develop and test to ensure that Sofia responds to customer queries the way they would like it to. Typical questions from HHWT users include, according to Goh: “What can I eat in Tokyo?” and “Show me South Korea itineraries.”

Currently, Sofia uses what is known as a decision tree that lets users narrow down the answers they are looking for, said Goh.  

“Sofia responds to queries by bringing users to the relevant content/articles available on the Have Halal, Will Travel site. With machine learning inbuilt within Sofia, it is able to deliver accurate answers to eight out of the ten questions posed to her,” he added.

But with its ability to learn about users’ queries, machine learning and larger datasets over time, the chatbot will be able to develop contextual and personalised content.

“As Sofia continues to take queries and respond to users, we are able to better understand what kind of questions our users have, and use this data to predict trends to uncover what people truly care about,” Goh said.

“We are also working on making Sofia smarter, to be able to add more value to users and travellers on our platform. For example, if I knew through Sofia that you are flying to Japan in two months or three months, then the next direct action it will take is to recommend products that will enhance your travel experience - walking down the street in a kimono, halal sushi-making classes and so on,” he added.

The company works with several regional tourism boards advising them on how to reach out and market their services and destinations to Muslims through content and campaigns.

In early May, HHWT announced it had signed a one-year partnership with the Hong Kong Tourism Board (HKTB) to produce content positioning the city as a preferred destination for Muslim travel. The pair will work together on travel guides, itineraries and tips on how to discover Hong Kong like a local while being able to access halal cuisine, prayer facilities and other Muslim-friendly amenities.

HHWT’s chatbot is available only in English, and on its desktop site.

(Reporting by Yosi Winosa; Editing by Emmy Abdul Alim

Our Standards: The Thomson Reuters Trust Principles

© 2018 All Rights Reserved