Exclusive: Inside ADB’s digital innovation sandbox programme

By Yun Xuan Poon

Interview with Ozzeir Khan, Director of Digital Innovation and Architecture Division, Asian Development Bank.

Just as toddlers learn how to swim in the wading pool, arm floats and adult supervision on hand, innovators often need a low-risk, well-regulated testing space whenever they float a new idea.

Tech companies call this an ‘innovation sandbox’. The Asian Development Bank’s (ADB) digital sandbox programme is less than a year old, but it is already making waves within the organisation.

GovInsider spoke to Ozzeir Khan, Director of Digital Innovation and Architecture Division at ADB, to understand the impact of the Digital Innovation Sandbox Programme, and how it plans to revamp ADB’s projects in the future.

Digital IDs for remote areas

The sandbox is useful for launching projects to help partner nations, and change internal operations. For instance, the sandbox is currently piloting a digital identity system in Papua New Guinea to promote financial inclusion. The villages had no formal system for identification, so people were not able to access financial services.

The sandbox team has programmed an Android device to issue identification numbers to villagers without an internet connection. Villagers receive an identification card once the machine has recorded their name and biometrics, and they can use this card to transact in the villages and be included in the financial schemes of the government, says Khan. ADB syncs data from the Android device to a central database every month or so, so there is a digital record of these villagers.

The local government is looking at offering healthcare benefits to citizens and validating their identity through these identification cards. Khan is hopeful that other service providers in the country will use the digitally-issued IDs to reach out to those people in the remote areas.

Revolutionising with AI


The sandbox is exploring and piloting three AI-based tools. One of these is a credit rating tool for smaller companies in developing nations. Khan’s team is working to come up with AI-based ratings for small companies using alternative sources of data, he says. This includes data from local sites within the developing nation and telco data. For these firms, “you will not find Moody’s and some of the rating agencies” to give credit scores and evaluates banks’ risk of investing in them, explains Khan.

Khan’s team is also trialing an AI-based feasibility study to assess the risk of new projects the ADB is embarking on. This tool brings together information on past projects done in a certain sector by ADB as well as other organisations like the World Bank; latest news of the country and the region; and the government’s future plans. “You bring the past, the present and future plans”, so that “AI is now recommending what needs to be done,” Khan says.

Third, the sandbox team has developed an AI-powered recruitment chatbot to shorten ADB’s hiring process. An application form can take a candidate up to three days to fill in, but this bot brings that down to 60 minutes. “The primary concern here is for the applicant to apply faster, and for [HR] to filter quickly so that we can get people on board faster, so we don't waste time,” says Khan.

The process is much faster because the bot fills in the application form for candidates by asking them questions specific to the role they are applying for. “It will dig in and try to ascertain your expertise”, says Khan.

The chatbot also answers applicants’ general questions about working in ADB. “It doesn't have to be job-related,” says Khan. “For example, if I work here for ten years will I get a pension, or what's the weather in Manila like?”

What’s next in the sandbox


Next, the sandbox team plans to create ADB’s very own ‘digital twin’ platform to help project managers make decisions about large-scale infrastructure projects, Khan says. This means they create digital copies of the projects to replace the huge physical models that were built in the past.

With a digital twin platform, engineers would be able to walk around a virtual road, tunnel, or bridge, and discover the possibilities. For instance, engineers building an airport would be able to study the effects that increasing its size would have on the traffic in the city or the environment. “You build this not just in 3D but in 7D,” says Khan.

This digital twin platform will pull in real data of the current population of the city, so engineers can study the true impact of any project on the people in the area. Project managers can also use the digital twin system to track the building progress. After the project is built, authorities can monitor the project to see if it has the expected impact on citizens.

Managing innovation and risk


Before deciding to develop any new idea, Khan’s team conducts a proof of value test to assess how realistic it would be to implement the project. The team asks questions like: does the project cost too much? Does it take too long? Does it fit in with ADB’s priorities?

This is on top of a proof of concept test, which just looks at how feasible it is to create a particular tech solution. “Technology-wise, you can probably prove many concepts. But will they work in our environment?” asks Khan. “Just because the proof of concept is true doesn't mean we will go into the creation stage without doing this sort of due diligence.”

The sandbox team embraces failure as part of their work, and they make sure that senior management understands this too. “We know that ten to 20 per cent [of our projects] are really moonshot, and chances of having a success are extremely low,” says Khan. “We develop this culture right at the beginning and walk [the senior management] through the chances of failure.”

For example, for the past six months, the team has been working on an AI algorithm to pinpoint the reasons behind the successes or failures of past projects. “We're still working on it. We're not a hundred percent sure, and there are times when it takes longer or it doesn't [work],” says Khan.

Supporting startups


Another priority for Khan’s team is to support startups armed with ideas and solutions to tackle climate change. Working towards the United Nations’ Sustainable Development Goals is a priority for ADB, and building up a community of environment-centered startups is an essential step in this journey.

“When I put up something like FinTech or payment, you will see thousands of startups interested, because there’s commercial value,” says Khan. “But when I look at agricultural tech, or ocean tech, or climate tech, the startups are not visible.” The sandbox programme will explore building a platform to support the growth of these startups so their solutions can reach a wider audience.

ADB has built a sandbox designed to hold that delicate balance between daring to experiment and minimising the potential negative impacts of a new project. It has already begun to transform workings within the ADB, and it plans to continue expanding ADB’s reach and impact.

They call it a sandbox but really, the concept is aquatic - they can easily test out whether a project will sink or swim.