When compared to industries such as defense, health care, and retail, innovation in the capital markets can seem slow. This year, though, the coronavirus pandemic forced workers around the globe to stay home, which put an incredible strain on IT teams at most every bank, asset manager, and vendor in the space.

As a result, automation stepped to the forefront. Firms relied on the cloud. They experimented with emerging technologies. And as employees shifted to working from home on the fly, financial institutions had to ensure that important tech projects that were already underway at the beginning of the year didn’t atrophy.

While these projects are not limited to any single region, firms in Asia had the advantage (or disadvantage, rather) of facing the shock and impact of the Covid-19 outbreak first.

The list of stories below is by no means an exhaustive list of all the innovation stories we’ve covered in the region. Instead, they are highlights of the lot and meant to serve as a glimpse into some of Asia’s more cutting-edge projects. For full stories, click on the links.

Think I’m missing something? Please let me know here: wei-shen.wong@infopro-digital.com. I’d love to hear from you. 

Deutsche Bank

Many of Deutsche Bank’s innovations this year sprouted from its Singapore base. Just before the coronavirus pandemic, DB had been ironing out unnecessary manual processes across the institution. 

Its efforts have helped the bank execute its business continuity plan better, said Stuart Gurr, group CIO for Asia-Pacific. Within two weeks at the end of February, DB enabled 50,000 people globally to work remotely with the videoconferencing capabilities on its internal platform.

He added that the bank focuses on what it calls “production engineering” to support the infrastructure, development, and technology teams. One of the tools it uses is intelligent virtual assistants (IVAs) that provide automated capabilities around the self-provisioning of products. DB’s IT service desk came under a lot of pressure as staff moved to remote working. The IVAs helped to direct traffic coming into the desk, as well as handling requests without the need for human intervention, thus freeing up human agents to focus on more complex requests.

In July, DB also announced it is partnering with Google Cloud, as it looks to redefine how it develops and offers its financial services. Chris Bezuidenhout, CIO for corporate and investment banking in Asia-Pacific and emerging markets at DB, said that the bank hopes to design new solutions with interoperability in mind. “We see opportunity in terms of challenging our design thinking more specifically, and working with Google to think about enabling more advanced interoperability as a theme,” he said. The focus in the beginning will be around advanced analytics, improving volume distribution and bank transaction processing, and strengthening its engineering acumen. 

Google can also help expand its use of NLP for its “Debbie” bot, which is designed to provide client trade status information and facilitate foreign-exchange workflows. 

One of the more interesting projects at the bank that incorporates Debbie involves its partnership with BNY Mellon. The two developed an API-enabled FX workflow to address the challenge of dealing in restricted currency markets, which often involved manual processes in performing validations and approvals before executing a trade. 

Darren Boulos, head of FX for Asia-Pacific at BNY, said that before entering into an FX transaction in those markets, the steps involved include the communicating of account balances, making sure trade details are accurate and ensuring that any income and corporate action components are approved and have settled. 

Gordon Alexander, head of client access and flow execution for Asia Pacific at Deutsche Bank, said helping international investors conduct FX conversions to buy a security can sometimes take as long as 36 hours to complete, as it involves satisfying various rules and approvals. Throw in the additional back-and-forth communication between operations, global custodians, and sub-custodians, and different time zones, and you have the perfect recipe for a long and drawn-out process.

The API-enabled FX workflow has been applied to custody of FX transactions in Korean won, and successfully reduced the pre-trade lifecycle from hours to seconds, Alexander said.

The solution uses chatbots built by DB and BNY—Debbie and Selina, respectively—which have since been integrated with collaboration services provider Symphony’s platform. Debbie and Selina were integrated in 2018 to facilitate the flow of information for clients trading securities on the Hong Kong Stock Exchange. 

Citi

Citibank is another firm working with Symphony. Its corporate venture arm, Citi Ventures, was completing work on a bot that helps traders deal with event data from sources that potentially affect commodities prices before traditional sources confirm it. 

The Price Shock Detection System (PSDS) bot was built by D10X, or Discover 10 Times, Citi Ventures’ internal program looking at new solutions for the bank and its customers. 

The bot was built during Symphony’s APAC hackathon in September, where it won in the “Work From Home Workflow” category. It will progress to the global Symphony Hackathon Olympics in 2021, competing against other contenders worldwide. 

As a start, the work on PSDS is focused on oil traders in the commodity space. It uses artificial intelligence to sort and classify signals on a variety of features to present traders with information before mainstream news sources write about those events. These sources can be from social media platforms, microblogs, as well as chat forums. 

“The question was, ‘How do we capture alternate sources of data and extract trends and developing events from them, and present them in a way that is actionable so that you get to be ahead of the curve long before someone confirms the news?’” said Agrim Singh, hacker-in-residence at Citi Ventures D10X in Asia-Pacific. “The added angle there was that this might be a rumor and there’s a reason [a major news outlet] hasn’t posted about it because they’re trying to confirm the facts. But markets may move regardless.” 

The D10X team used Symphony’s platform to request the stream of notifications or subscribe in a way that every few minutes, as events develop and more information becomes available, they can push the notification out to traders. In a day, they built a rudimentary version of the bot. 

Singh and the team are working to incorporate Citi’s data and build the bot out. 

After that, it will go through testing with traders and get feedback. Once it’s in production with the oil trading team, only then will D10X look to scale it to other desks and traders. 

Symphony

Symphony Communications integrated support for more text messaging applications such as LINE, which is widely used in Japan, and Apple iMessage. These are additional options to its existing portfolio of messaging applications, which allows financial firms to onboard customers and interact with them on these platforms. 

Symphony launched WhatsApp Connect earlier in April, and before that, in December 2019, it launched a connectivity channel for WeChat, the Chinese social media and messaging platform. David Gurle, CEO at Symphony, told WatersTechnology that Symphony is developing additional functionality for all its WhatsApp and WeChat integrations. 

This will allow clients to exchange files, such as research documents or trade confirmations. The individual messaging channels are encrypted to ensure the secure transfer of information between counterparties. All communications and activities are also monitored to comply with regulations like Mifid II. They are then exported to the institution’s surveillance systems and checked by the compliance teams if needed.  

“You can continue to send messages to an external party on Symphony with all the compliance capabilities and information abilities that are the discretion of the institution for which you work, and all the messages are logged and tagged to distinguish them from Symphony messages, WhatsApp, or WeChat,” Gurle said.  

Symphony’s WeChat integration differs from the other messaging channels. It had to create a mini app on Tencent’s app store (Tencent owns WeChat). Users can onboard counterparties to the WeChat stream using a QR code. 

Deutsche Bank, for example, is using WeChat on its Symphony platform to allow corporate and institutional clients to complete multiple steps in the FX trade process. 

Finos

The Fintech Open Source Foundation (Finos), a non-profit organization promoting open-source development, has built its name within the investment banking space. It has relationships with Citi, Deutsche Bank, Goldman Sachs, JP Morgan, Nomura, RBS, UBS, and many more. 

But that’s not enough for Finos. It wants to expand its reach to retail banking too. Gabriele Columbro, founder and executive director of Finos, told WatersTechnology that the firm’s ultimate goal is to host open-source or open-standards collaborative efforts that solve industry-wide business challenges in the retail space. 

It has two approaches to help it inch towards its goal of going “inside-out” and “top-down.” Using the inside-out approach, Finos will use existing technology within its member and contributor base, of which it has over 300 contributors in the capital markets. 

Meanwhile, the top-down approach means that Finos will turn to existing open collaborations such as open banking, or regulations like the revised Payment Services Directive (PSD2), and evaluate where open-source projects can accelerate standardization and implementation processes.  

Closely intertwined with its retail plans is the Asia region. According to Columbro, the line separating incumbents from start-up fintechs is “more blurred” there than in the US or Europe. 

Since April 2020, Finos has been working with the Linux Foundation, which Columbro said has a very developed organization in Asia, especially in China, Korea, and Japan. “This is just one more step towards what we think Finos can become—the worldwide umbrella for any type of open-source and open-standard collaboration in financial services,” he said. 

ASX

The Australian Securities Exchange (ASX) is well-known for being one of the first to take the plunge with distributed-ledger technology (DLT) with its ambitious CHESS replacement project. While it has yet to successfully roll that out, in part due to Covid-19—which has extended its implementation timeline—the ASX has other projects up its sleeve. 

In September, it opened up DataSphere to third parties looking to partner with the exchange to solve data challenges. Since launching in late 2019, the data science platform has added new datasets, including data from outside entities. 

David Raper, executive general manager for ASX’s trading services, said the exchange is trying to figure out how to capture and monetize more of the data it creates. “ASX makes a lot of data available today, so when you see the news that BHP Group has been up 3%, for example, that’s from ASX data. But there’s a lot of other data that is generated within the organization that has never been surfaced,” he said.

The exchange is working on a range of products using DataSphere to allow traders, banks, brokers, asset managers, and custodians to understand the bond, money, and repo markets better. Data in these markets are traditionally opaque, he added. The platform will help make these markets more transparent and will help participants understand the volatility of bonds and money market instruments and the liquidity and concentration risks associated with particular securities. 

ASX DataSphere has two workspaces—business and data science. Business workspaces allow users to analyze datasets through visualizations and spreadsheets, where participants can also upload private data. 

HKEX

Hong Kong Exchanges and Clearing Ltd (HKEX) is exploring opportunities to monetize the data it produces. As an example, it is creating a scalable Data Marketplace platform, which will allow the sharing of data and analytics through a commercial pricing mechanism. 

However, it seemed that the exchange is still forming its strategy on how it will execute plans to do that. 
In 2019, HKEX bought a minority stake in Beijing-based data technology company, Huakong TsingJiao Information Science (Beijing) Ltd, which HKEX believes will play a key part in helping develop its future data marketplace.

TsingJiao specializes in multi-party computation technologies—a subfield of cryptography—that allows for collaborative data analysis without revealing private data during the analysis process.  

Charles Li, HKEX’s outgoing chief executive, said in a press briefing that HKEX’s role could be in the organization, development, and implementation of a technology platform that solves the problem of data privacy and data protection. “We are constantly looking at that particular role that we are very good at, which is connecting the dots and bringing everybody together, building up our alliances so we can start to build out our ecosystem in data, and that’s where we put the investment in,” he said.

HKEX has also bought a 51% stake in Shenzhen-based Ronghui Tongjin—now known as BayConnect—which specializes in regulatory and exchange technologies. It has also signed a memorandum of understanding with Ping An to explore potential collaboration areas in fintech, artificial intelligence, and data analytics to support the mutual connectivity of the mainland Chinese, Hong Kong, and international markets. 

Ping An

The battle for AI supremacy may yet be called, but China is showing off its advancements, particularly in the field of natural language processing and generation. 

Although not a household name in Western nations, Ping An, the Shenzhen-based insurance, banking, and investment giant, is consistently ranked at or near the top of the General Language Understanding Evaluation benchmark—a list that often includes the likes of Google, Alibaba, Huawei, Microsoft, and Facebook. 

Capital markets firms are increasingly leaning on NLP for chatbots to extract information out of dense, unstructured reports. We at WatersTechnology have written a lot about NLP this year. 

Ping An’s Smart Audio Robot was built in two days using the firm’s Omni-Sinitic core framework to help contain the spread of Covid-19 in Wuhan, China. The robot was equipped with investigation functions, follow-up alerts, and the ability to send reminders automatically. It made daily phone calls to gather information on residents’ symptoms and body temperatures, allowing epidemic prevention personnel to focus on more important tasks.  

Although this particular use case first serves the consumer market (and government), as it further develops its core framework, the use cases will naturally extend into the wholesale and capital markets.

Jing Xiao, chief scientist at Ping An, said the firm also applied the Omni-Sinitic framework within its investment research and risk management processes.

JP Morgan Asset Management

In line with NLP technologies, JP Morgan Asset Management (JPMAM) is building out its Textual Analytics tool to read Chinese and Japanese documents. 

The tool reads millions of documents, ranging from company filings and corporate event transcripts, to employee reviews on sites such as Glassdoor. Currently, the tool has only been trained to read English documents. 

Over the summer, Will Coulby, global equity data science lead at JPMAM, told WatersTechnology that the team is debating whether it makes more sense to add language variance to the model, or to translate other languages into English before putting them through the model. 

“At the moment, we’re tending towards the translation, just because the scale of the documents we’ve got seems to work a bit better through this model. The core of this type of approach is that the model needs to have to read a lot of information to gain value from it,” he said. 

Besides potentially adding new language variants to the model, JPMAM also considered new datasets on which to train the model.

Although Textual Analytics is now used by its internal portfolio managers, JPMAM could commercialize it later. Both its quantitative and fundamental investment teams use the tool. 

Over time, JPMAM will add more data, topics, and predictions to the model. “With every new piece of feedback, the model is able to better understand this data. Having an ongoing team [providing] specialist, targeted feedback gives us a real competitive advantage in model training,” he said. “More importantly, though, it builds upon our approach of the model trying to read content like our analysts do.”