The accessibility of cloud-based computing power and high-performance networking, the availability of increasingly large and varied datasets, and a wealth of readily obtainable tools for building models to analyse the data - have all led to widespread adoption of AI and ML in the financial markets industry.
And as the technology continues to advance, new use cases are constantly emerging, from the front-office to the back, across the full spectrum of pre-trade to post-trade, including portfolio management, trade execution, risk, regulation, and compliance.
However, in order to support the unique requirements of AI and ML applications in financial markets, a robust networking and connectivity infrastructure is essential. As firms increasingly rely on these technologies, they need to know they can access the necessary data and the associated tools and services securely, reliably and at speed.
AI use cases in financial markets
For the buy side, AI and ML – use models and algorithms that learn from data to make decisions without being explicitly programmed. They’re now changing the way investment decisions are being made. ML-based models are now often used in pre-trade analysis. High-frequency trading firms, systematic traders, quantitative hedge funds, and others – analyse signals in market data to identify trading and investment opportunities and uncover alpha.
One of the key advantages of ML is its ability to extract information efficiently from a wide range of large datasets, whether numerical, textual, structured or unstructured. Natural language processing techniques are particularly useful here, as they can extract meaningful information from various sources of unstructured text. And while few funds have developed a fully end-to-end AI-based investment process, hedge funds run by analysts with strong ML backgrounds are becoming increasingly widespread.
AI also has a range of use cases at sell-side firms, such as improving trade execution performance; optimising hedging and quoting decisions; and bringing about workflow efficiencies (by automating brokers' responses to client requests for quotes, for example). A recent report from ESMA stated that AI finds some of its most promising applications in the trade execution phase. They highlighted the example of investment banks and brokers using AI-driven execution models to intelligently split large parent orders into multiple child orders across trading venues and over time periods, to minimise market impact and reduce transaction costs.
From a post-trade perspective, although AI adoption is still nascent because most central securities depositories and central counterparties rely on legacy technology infrastructures. Some of these institutions are now investigating how supervised and unsupervised ML techniques can help with trade settlement, clearing, and reporting. For example, some data reporting service providers and trade repositories are now developing AI solutions such as anomaly detection and automated data extraction from unstructured documents, with the aim of improving the efficiency and accuracy of post-trade processes.
Regulators are also making increasing use of AI and ML, particularly in fraud detection and prevention, where large quantities of data can be rapidly analysed to identify patterns of suspicious behaviour and highlight potentially fraudulent activities. Stock exchanges and other trading venues are also applying AI to detect irregular and potentially malicious trading activity. Nasdaq, for example, now leverages specific machine learning capabilities for market surveillance, including deep learning, transfer learning and ‘human-in-the-loop’ learning.
AI and ML applications that require real-time data to be processed and analysed require fast, reliable network connectivity. And because they rely heavily on large data sets - which must be stored and processed reliably and securely – they’re often built and run on the cloud, not just for scalability, but also for the wealth of cloud-native AI and ML resources that are now available, such as Google TensorFlow, AWS Machine Learning and Microsoft Azure AI.
For firms looking to utilise such cloud-based applications within financial markets, security is a top priority when it comes to networking and connectivity. Firms need to know that their sensitive financial data is being transmitted and stored securely, with advanced encryption techniques where necessary, to ensure that the risks of unauthorised access, data breaches and cyber attacks are kept to a minimum.
How BT Radianz helps
By providing access over a single connection to more than 400 third-party technology providers with thousands of applications and services, the BT Radianz cloud ecosystem has become a key component in the growth of AI and ML usage in the financial markets sector.
We enable banks, brokers, trading and investment firms, exchanges, trading venues, and clearing houses to leverage those AI/ML applications and services across a wide variety of use cases, across the front, middle and back office, from pre-trade to post-trade, in functional areas including portfolio construction, decision support, trade execution, risk, regulatory compliance, clearing and settlement.
Getting ready for DORA
As regulators strengthen the financial system’s ability to absorb events such as pandemics and cyberattacks - resilience is under the spotlight.
The cost of GPS vulnerability
Without a secure backup, a GPS outage could put your organisation at risk of disruption, losses and outages.