There's no doubt Government websites can be made more personal.

There's a trove of information in each portal, but the information is under many layers, and such an oversupply of information can be daunting for citizens who get frustrated with trying to navigate around the sites.

Some users may prefer chatting with 'someone' as a way to locate the information, while there are users who may not like to, or cannot use, a keyboard.

To push the development of virtual assistants that will add a touch of personality to Government websites, the Government Technology Agency (GovTech) and Microsoft co-organised the inaugural Botfest 2016.


The audience at Botfest 2016 enjoying the chatbot demonstrations by the competing teams.

Exclusively designed for government officers, it provided an avenue for agencies to use artificial intelligence (AI) and machine learning to solve real challenges in their domain areas.

By developing these chatbots, as they are popularly called, government agencies are thus seeking to make it even easier for citizens to access an extensive range of online services, from applying for parking permits to budgeting for hospital fees.

In the first part of Botfest, nine teams of participants underwent a training session on 31 October at Microsoft Singapore's Operations office. Following that, the teams had 4 days to build chatbots using the Microsoft development environment and Microsoft Bot Framework.

11 November was D-Day: All teams were invited to present their projects.

Dealing with the Deluge of Online Info


SPRING Singapore’s web chat application, specially designed to help entrepreneurs, won first prize at Botfest 2016.

Springing into first place was SPRING Singapore’s web chat application, which facilitates high quality entrepreneurial linkages by increasing opportunities for productive partnerships and availability of networks in Singapore tech startup eco-system.

It starts off as a chat screen to prompt queries, and responds by launching a companion web browser screen that displays the relevant information for the end user.

It won over the judging panel comprising of Mr. Tan Kok Yam, Head, Smart Nation Programme Office (SNPO); Mr. Richard Koh, Chief Technology Officer, Microsoft Singapore; and Mr Dennis Lui, Assistant Commissioner (Taxpayer Services Division), Inland Revenue Authority of Singapore (IRAS).

The Ministry of Health’s (MOH) chatbot, which was the runner-up, provides information on procedures and other considerations that would determine their medical fees at both private and public hospitals.

In third place was the Singapore Examination and Assessment Board’s chatbot, which enables pupils to pull publicly-available information about exam dates, times, syllabi and venues, as well as the distance from their current location to the exam venue.

MOH team member Mr. Terry Lim said that his team’s motivation for their chatbot was the “overwhelming” information available on MOH’s website.


The friendly chatbot from the MOH team helps users to navigate 'overwhelming' information even medical terms to determine their likely fees at both private and public hospitals.

“While an Internet search may provide links to various tables of procedures and costs from the MOH website that users have to figure out, the bot was built first to understand layman term equivalents of medical terms such as colonoscopy and cataract surgery,” Mr. Lim explained.

“Thereafter, the bot will not just offer the relevant costs that are directly applicable, but will also go further to help users understand the options and allow them to compare costs based on their preferences in a Q&A conversation. Where necessary, the bot would even provide relevant information from other websites.”

During the design stage, one of the main challenges and source of frustration the MOH team encountered was dealing with free-form responses from users, Mr. Lim recalled.

This was dependent on the maturity of the Natural Language Processing algorithm of the underlying bot framework to recognize the intent of the user’s question, and thereafter, how much bot training was required, using possible variations of how questions could be asked.

Moving forward, Mr. Lim said the team aims to “close the conversation” such as helping users to book appointments with shortlisted hospitals where such APIs (application programming interfaces) are available.

With increasing accuracy in voice recognition, the team would also consider incorporating voice inputs and language recognition for dialects when the tech is ready.

Reducing Pain Points

A GovTech team built a chatbot to help answer questions about grant applications on the Business Grants portal.

Besides improving the efficiency of applications, GovTech’s Grant chatbot sought to help applicants find the right grants to apply for their businesses, reducing the time and resources required by administrators to manage grants.


The GovTech chatbot helps to answer questions about grant applications on the Business Grants Portal (Portal) and has an OCR function.

“Working on the Business Grants Portal, we spent a lot of time thinking about how to make it easier for people to complete and submit forms,” said Ms. Nina Ee from team GovTech.

“We started thinking if a bot could help clarify questions during application, make the process smoother when an applicant needs to provide additional information and allow applicants to submit documents, these features would help facilitate the submission process and save time and effort for the user.”  

An obstacle the team encountered was how best to tap into optical character recognition (OCR) to identify information from paper invoices, which is then read into an electronic format.

“Many people have tried OCR and found that this technology was challenging to use as the information gathered is not fully accurate. To overcome this, we had to design and come up with our own algorithm to work around the problem,” she shared.

Government Services, Galore

Other entries came from the Maritime and Port Authority of Singapore (MPA), whose ChatBot is designed to generate interest into maritime careers, provide social listening and reach out to citizens to join the maritime industry.

The Ministry of Communications & Information developed Moli, short for Moments of Life, which will keep users abreast of policies or schemes that are applicable to each stage of a Singaporean’s life.


The HDB team behind the Parking Advisor chatbot, or Parking Auntie as we like to call 'her'.

The Housing and Development Board’s (HDB) Parking Advisor, called Parking Auntie, provides a one-stop hassle-free service to help citizens apply, renew and pay for season parking, find nearby car park lots, and answer questions on season parking.

Academic team Temasek Polytechnic developed a chatbot that will answer Student Services enquiries.

But the crowd favourite was the Defence Science and Technology Agency’s My NSBuddy, which the team presented with a short skit.

My NS Buddy answers questions relating to National Service, connects Servicemen with each other, and even provides guidance and motivation to active National Servicemen.

GovTech co-organiser Ms Gladys Tay said: “We were pleasantly surprised to receive strong response from the agencies to participate in the Botfest. It provided an excellent opportunity to accelerate learning and getting agencies to start thinking about using conversations to interact with citizens."

The friendly neighbourhood chatbot has thus become an additional channel that helps Government agencies to engage and deliver digital services to the citizens.

Chief Executive of GovTech, Ms Jacqueline Poh said, “Botfest has generated many creative, interesting and practical use cases that will enrich and provide additional insights to our chatbot exploration efforts.”