Leveraging tech to make a quantum leap

Once considered a niche research area, quantum computing may find its way into financial institutions sooner than we think.

Many experts predict seeing a ful­ly fault-tolerant quantum computer by 2035. Should that happen, McKinsey estimates quan­tum computing could create US$622 billion ($798 billion) in value for the finance indus­try if utilised to improve existing processes and for transformative use cases.

“Due to rapid tech advancements, it may not be long till quantum computing becomes mainstream, marking a paradigm shift from classical or conventional computing. OCBC has been taking various steps to ensure we’re ready to capitalise on that shift,” says Praveen Raina, head of group operations and tech­nology at Oversea-Chinese Banking Corpo­ration (OCBC).

Safeguarding the bank and its customers

Quantum computing is believed to be revo­lutionary. It uses the laws of quantum me­chanics to explore many paths in parallel through superposition, enabling it to deliv­er a range of possible answers faster than a classical computer.

Quantum computing could revolutionise portfolio and risk management. For example, quantum-optimised loan portfolios focusing on collateral could allow lenders to improve their offerings, potentially leading to lower interest rates and freed-up capital. Beyond commonly mentioned use cases like portfo­lio optimisation, quantum computing could revolutionise areas such as trade finance and collateral optimisation in corporate banking due to the high monetary value and com­plexity involved.

See also: Navigating the workforce shift for a GenAI-powered future

Similarly, they also have the potential to break cryptographic encryption, which pos­es a major cybersecurity concern. “Quantum computers can break those keys easily due to their superior computational power. This is why we’re actively exploring the appli­cations of quantum computing in the area of cryptography to safeguard customer data against imminent future quantum attacks,” states Raina.

Since the start of this year, OCBC has been working on Quantum Security solutions, such as Post Quantum Cryptography (PQC) and Quantum Key Distribution (QKD), with vari­ous ecosystem partners such as institutions of higher learning, regulators and private sectors.

For instance, OCBC is collaborating with the Monetary Authority of Singapore (MAS) and other banks to experiment with QKD solu­tions. The collaboration supports the advisory MAS issued in February 2024 that warns finan­cial institutions about cybersecurity risks as­sociated with quantum technology and offers guidance on mitigating the identified risks.

See also: Updated version of Singapore's unified QR code, SGQR+, launched by NETS

PQC and QKD are currently the primary methods for quantum safe communication to elevate cyber defence posture. PQC algo­rithms are classical cryptographic techniques designed to be computationally challenging for quantum computers to crack. QKD is hard­ware-focused and uses quantum mechanics principles to establish secure communication channels and ensure no interceptions of the exchange of encryption keys or eavesdropping.

Raina shares that beyond cybersecurity, OCBC is also experimenting with quantum computing to elevate risk management and derivative pricing, among other use cases.

Supercharging AI

The convergence of quantum computing and artificial intelligence (AI) will lead to trans­formational changes. There is a symbiotic re­lationship between quantum computing and AI, whereby quantum computing can rapidly process extensive data sets, driving enhanced computational power, parallel processing, and more efficient problem-solving. It can also be leveraged for quantum machine learning (QML), accelerating AI’s learning processes. Similarly, AI can be also used to drive en­hancements in quantum computing through error correction, noise reduction, quantum al­gorithm design and optimisation, simulation and emulation. OCBC hopes quantum com­puting will further improve AI capabilities to enhance operational efficiency and contribute to higher levels of customer service.

“AI is not new to us; we’ve been on our AI journey since 2016. But with the advent of generative AI (GenAI), we have been push­ing boundaries further. Today, we’re running more than 300 AI models offering personalised financial advice, providing seamless and in­tuitive customer experiences, and optimising operational efficiency,” states Raina.

More than five million decisions – in the areas of risk management, customer service and sales – are currently made by AI in OCBC daily. This is projected to increase to 10 mil­lion by 2025. For instance, the bank has AI models that quickly analyse customer profiles and networks to identify customers who are more susceptible to scams.

“[Recognising its importance,] OCBC has a governance structure to scale AI without stifling innovation so that employees can continue experimenting with AI in various ways,” says Raina.

Whilst the utility and benefits of AI are ex­tensive, there is also a need to be aware of the potential risks associated with it. The issues are well-documented: ‘hallucinations’ (GenAI models providing inaccurate answers), ethi­cal/reputational risks, risks of promoting ine­quality, ascribing intentions, and data security frauds, to name a few. Addressing concerns around AI ethics, OCBC has put up guard­rails to prevent GenAI from generating false information. This includes having its own LLM evaluation framework to test models be­fore putting them into production and using large language models (LLMs) to check other LLMs to ensure the accuracy of their results.

Building a quantum-ready workforce

To start tapping and extracting the full value of technologies such as quantum or AI, tal­ent in these areas needs to be built.

According to LinkedIn’s data, the compe­tencies required for all jobs will change by 68% by 2030 due to the rapid rise of AI. To help its employees remain relevant, OCBC an­nounced in 2023 that it would invest $30 mil­lion over three years in skills development, which builds on the bank’s earlier $50 mil­lion investment from 2018 to 2023.

Initiatives targeted at staff include an in­ternal coaching certification programme, skills portfolio workshops, and a Data Cer­tification Pathway to groom data profession­als in-house through 12-month to 18-month training. Moreover, OCBC’s employees can leverage an AI-powered career marketplace to assess their current skills, explore inter­nal job opportunities, and identify the skills needed to advance within the bank.

Similarly, OCBC has taken the initiative to upskill staff on quantum through structured training programmes in partnership with in­stitutions of higher learning and the private sector. Over the next two to three years, the bank targets tripling the number of staff with an intermediate proficiency level in quantum technology.

To deliver a truly seamless and intuitive customer experience, OCBC aims to put a life­like AI voice assistant, like in the movie Her, in every customer’s pocket through its mo­bile banking app. “This is the common vision for the moonshot AI projects we’re working on. It’s a big goal, so we’re breaking it into small steps and getting more employees in­volved because we believe a bottom-up ap­proach fosters better innovation and employ­ee engagement than a top-down approach,” Raina shares.

One way the bank is encouraging bottom-up innovation is through the quarterly “Deal or No Deal” innovation contest. Employees can form groups (even with colleagues from a different department or country and even those without any technology background) and submit their ideas on how AI and other emerging technologies can help improve an existing process or tackle an issue in a new way. A panel of internal judges will then de­cide on the best idea to be developed and put into production.

Launched in January 2023, close to 60 ideas have been collected from employees. Some of the ideas received include an auto­mated portfolio review tool that can save over 80,000 man-hours a year, a hyper personalised commentary generated by LLMs, and smart lighting focused on cutting carbon emissions through IoT-enabled devices.

Raina adds that the bank’s IT staff gets one day a month to learn new technical or specialised skills. He explains: “This gives them dedicated time for training and empow­ers them to own their skills development. We’ll provide the licences to tech tools and online learning platforms like Udemy so they can choose what they want to learn. To en­sure accountability, their manager will have visibility on the topic they’ve learnt, and they are expected to apply the new skills to their work.”

SIDE STORIES

Future-proofing the bank in a sustainable way

OCBC anticipates heavy usage of pow­er-hungry AI and quantum comput­ing in its operations and is constantly exploring ways to do so more sustain­ably. Currently, its data cen­tre in Singapore not only re­ceived the BCA-IMDA Green Mark Platinum Award but is also 30% more energy-effi­cient, emitting 9.6 million kg less CO2 emissions than the standard data centre in Southeast Asia.

The bank, Raina adds, is partnering with like-mind­ed institutions for quantum computing as it is still a nas­cent area. “We have an inter­nal task force that focuses solely on quantum comput­ing, and they are prototyping with different organisations (including other financial in­stitutions), the government and universities [to tap on collective intelligence].”

Singapore’s government will also invest close to $300 million over the next five years to fuel quantum technology research and talent develop­ment. Specifically, this fund­ing will focus on four areas: scientific studies, talent, en­terprise, and engineering ca­pabilities. This is very much in line with global trends of public investment in quan­tum computing, which to­talled around US$42 billion as of 2023 and is only set to continue growing.

To further cement its com­mitment to innovation and customer-centricity, OCBC will invest about $500 million in the Punggol Digital District. The funds will be used for a strategic partnership with the Singapore Institute of Tech­nology and a new innova­tion hub expected to be com­pleted in the first quarter of 2027. Called OCBC Punggol, the 430,000 sq ft innovation hub will house up to 4,000 employees, with the majori­ty from the tech teams work­ing on emerging technologies such as quantum computing and mixed reality.

Some ways OCBC is leveraging generative AI to improve productivity and efficiency

• OCBC Wingman — A coding as­sistant that automatically gener­ates, debugs and optimises com­puter code to accelerate software development times by 20%.
• OCBC Whisper — Speech-to-text tool that can analyse all sales calls with customers to automatically identify potential anomalies in the sales process. The bank’s contact centre also leverages Whisper to transcribe and summarise calls in near re­al-time.
• OCBC Buddy — A GenAI pow­ered knowledge assistant chat­bot that helps employees get quick answers on OCBC infor­mation, such as company poli­cies and procedures, acronyms and annual leave.
• Document AI — Extracts and summarises key information from lengthy documents (like finan­cial and sustainability reports) in seconds through a drag-and-drop process.

The bank is also using other forms of AI to deliver more val­ue to its customers. One exam­ple is OCBC Securities’ Customer Artificial Intelligence Radar (A.I. Oscar), which can predict stock price movements and generate hyper-personalised stock ideas.

A.I. Oscar uses deep learn­ing algorithms to identify and predict market patterns to thor­oughly understand an investor’s trading patterns. It analyses fac­tors such as the individual’s risk appetite, past trading activity, and demographic information to curate and produce a list of 15 hyper-personalised stocks every week from exchanges in Singapore, Hong Kong, and the US. This allows investors to fil­ter stocks that are highly likely to meet their trading preferences and criteria and immediately act on the idea by making a trade.

Photo: OCBC Securities

According to OCBC Securities, trading activity among young investors under 35 increased by 50% during the pilot period of A.I. Oscar in 2023. It aims to triple the total number of active customers in this segment over the next three years by leveraging A.I. Oscar’s unique proposition.