by Zafar Anjum

How ADB is future-proofing itself with digital innovation sandboxes

Aug 04, 2021
Artificial IntelligenceDigital TransformationFinancial Services Industry

CIO Shirin Hamid and sandbox programme director Ozzeir Khan describe the bank’s approach to digital innovation—and the lessons learned along the way.

Nearly six years ago, Shirin Hamid joined the Asian Development Bank (ADB) as its director general (the CIO position). The ADB is a regional development bank established in 1966, headquartered in the Philippines, committed to achieving a prosperous, inclusive, resilient, and sustainable Asia-Pacific region, including efforts to eradicate extreme poverty.

Before joining the bank, Hamid was with the United Nations Development Programme for 11 years. Originally hailing from Singapore, she loves development work and is cognizant of how IT can bring impact and make a difference in the lives of people.

shirin hamid adb square ADB

Shirin Hamid, director general (CIO), Asian Development Bank

Shortly after Hamid joined ADB, the president of the bank talked to her about the five drivers in the modernization and reform of IT in the bank. Within six months, she and her team managed to create a strategy for ADB IT reforms, with 14 programmes. The team also put together 10 business cases, subsequently followed by four additional business cases.

Launching ADB’s digital innovation sandboxes

Digital Agenda 2030 is the digital blueprint that provides the vision and road map for ADB’s digital transformation through the use of secure, modern information and technology systems and digital processes to enhance effectiveness, efficiency, and resilience.

In 2017, Hamid started working on this digital agenda, aligning the bank’s IT strategy with its overall strategy. She embarked on a trimodal approach to tackle this new challenge: keeping the lights on, growing the organization, and future-proofing the bank through six programmes focusing on operations; financial services, administrative and corporate services; digital workplace and connected data; enabling the digital backbone; and the digital innovation sandbox, she says.

As part of the future-proofing digital agenda of the bank, she set up technology-specific sandboxes at the bank, primarily to drive digital innovation in a systematic and safe way. The peripheral drivers behind this move were to support digital transformation, to ensure that the bank’s technology is always fresh and technology debt is low, and to experiment with emerging technologies for the digital workforce such as robotic process automation (RPA), artificial intelligence, chatbots, and distributed ledger technology such as blockchain. Adoption of these new technologies would lead to a shift in the digital mindset of the stakeholders throughout the bank.

“The sandbox provides a safe space for risk-averse organisations like ADB to understand new technologies and envision potential benefits to the bank,” Hamid says. People embrace the technologies and step up as champions when they understand and therefore have the confidence to use them.

“Experimentation within the safe space of the sandbox involves real-life transactions. Working on the trade finance blockchain for the bank, for instance, opened doors for the business to understand how a digital consortium in a digital ecosystem works,” she says.

ozzeir khan adb ADB

Ozzeir Khan, director of the Digital Innovation Sandbox Programme, Asian Development Bank

Work on setting up the sandboxes started in 2019 with the intention of understanding the new technologies, recalls Ozzeir Khan, director of the Digital Innovation Sandbox Programme at ADB. Under the sandbox programme, 13 initiatives were piloted from October 2019 to January 2021, with the aim to develop them into full-scale digital products to be made operational over the next year.

Some of the programmes included creation of MT910 receipt records for loan-service payments, digitizing the non-SWIFT bank statements (extracting key information from bank statements), and country priority assessments using RPA; developing an AI-based system called Eva to scan thousands of evaluation documents and identify useful lessons in a dashboard categorized around country, sector, theme, modality, and year; and deploying an HR chatbot called MyADB Recruitment Intelligence (Mari), an AI-powered platform for creating chatbots to conduct competency-based structured interviews. The bot automates the candidate screening and qualification workflow.

Similarly, the team used distribnuted ledger technology to record transactions between two or multiple parties efficiently and in a verifiable and permanent way. Earlier, the processes involved highly manual processes, from the applicant’s email request to manual encoding of guarantee data in Trade Finance Processing Management System and manual review of data accuracy.

The challenges of scaling up the digital innovation sandboxes

There was a lot of success in early adoption of the sandboxing programme. But there were challenges, too.

First, there was risk aversion; mature organisations are historically averse to taking risks and can refuse to understand new technologies. Second, the new model required “co-creation of requirements and solutions, developing partnerships, working with the ecosystem of startups and supply chains,” Khan says. But that was a new way of working with the business and providers.

“Once the business understood the technology, adoption became quick and easy,” Hamid says. “The first crucial step towards adoption is to help them understand the technology and see the opportunities it will bring.”

Ironically, the early success of this sandbox approach threw a new challenge to Hamid’s team: Suddenly the volume of demand increased and pressure built up on the team to scale up the pilots. “Because the bank has seen the value and impact of technologies (like RPA and bots), the challenge for us now is how to keep up with the demand—both for pilots and for scaling,” she says. “This is challenging from a capacity point of view.”

Hamid’s dilemma was how to choose which projects to ramp up and give priority to and which ones to relegate to the waiting queue. On top of that, the COVID-19 pandemic made automation and remote work key priorities that had to be addressed regardless of other priorities.  “Demand for innovation exploded at the institutional level for the bank to survive and thrive,” she says. “The Philippines, where the ADB HQ is located, is experiencing one of the longest lockdowns in the world—further highlighting the need for innovation to support extended remote-work arrangements.”

What worked and what didn’t in the digital-sandbox experiments

As the projects matured, it also became clear that all the new technologies didn’t work equally well. “The adoption of a technology depends on the timing of the maturity of that technology,” Khan says. For example, 3D printing didn’t have much uptake. Similarly, the augmented reality (AR) and virtual reality (VR) technologies haven’t matured and could not support any meaningful implementation.

In Khan’s eyes, tokenisation is an interesting example of a technology that didn’t seem to be hot when it appeared on the horizon but then it suddenly assumed a new significance. “We thought tokenisation will be a very valuable technology but we had it on the back burner for two years,” he recalls. In 2020, digital currency started becoming mainstream and tokenisation technology became hot again.

Among the technologies that received most traction were AI, RPA, and chatbots. “The technologies that have been successful for us, and we believe will continue to be relevant, are the typical tech seen in the financial industry: using AI for collecting data for lessons learned and insights for future planning, big data and computing power, and automation of back-end financial services (via RPA and chatbots),” Khan says.

“The AI part caught us by surprise,” Khan says. “If you let AI loose, it gives you a lot of insights.” Now, the ADB team is using AI for planning. For example, they are using AI in predicting the GDP rates. AI is also helping the bank decide where to invest more to impact development.

Going forward, ADB is focusing on those technologies that are growing fast such as AI and machine learning, especially natural language processing (NLP); tokenization; digital twins for designing infrastructure projects; and internet of things and drones for monitoring projects. Technologies that are in nascent phases of development include synthetic biology and data.

“As we plan for our new operating model, post pandemic, we are taking the lessons we’ve learned in the past year and a half to enable the technologies that will support the new normal,” Hamid says. “The challenge now is transitioning to infuse digital innovation into ADB’s day-to-day practices and moving forward by taking into consideration organizational, regional, and global drivers,” from bank culture through the structural and workplace changes wrought by COVID-19 pandemic.