Fintech : AI & machine learning : Jannah Patchay

THE ART OF DELEGATING TO TECHNOLOGY.

Jannah Patchay gets behind the jargon and looks at how new technology is impacting the trading and investment world.

Artificial intelligence (AI) and machine learning (ML) are perhaps two of the most misunderstood and misused terms in the modern lexicon. These technologies may be slowly revolutionising the way in which trading venues and participants operate but the death of the sales-client relationship or an increasingly disconnected market landscape, are largely unfounded.

At the moment, the technologies are mainly targeted at data management and insights. Although often labelled as separate functions, ML is in fact but one application of AI, in which computers are programmed with algorithms that enable them to learn and adapt, and are given vast amounts of data on which to ‘train’. It is typically used to collate vast amounts of market data, determine market behaviour and train algorithms to recognise patterns of behaviour. Trades are then executed based on their assessment of current market conditions.

AI, on the other hand, is an umbrella term. It is a wide-ranging concept involving the ability of computers to execute tasks and make decisions in a way that exhibits characteristics of learning and intelligence, as opposed to algorithmically pre-defined programming logic.

Different solutions with a twist

Not surprisingly, there are many different products on the market. For example, Tyler Capital, a proprietary trading firm and liquidity provider has rolled out trading system OpusOpus, a dedicated and highly specialised ML-based application. The aim, according to Chris Donnan, Tyler Capital’s CTO, is “the systematic application of machine learning to the global financial markets, with our priority being to operate a process that can consistently create safe, scalable, repeatable, reliable goal-directed adaptive systems that operate in many environments.”

Tyler Capital operates at one, highly automated end of the spectrum, that deploys ML in sales and trading environments. However, there is another, potentially much larger group of market participants who think that the technology’s key benefits lie in its ability to reinforce sales-client relationships, and to improve the firms’ offerings to their clients.

 

 

For example, at NatWest Markets, the investment banking arm of RBS Group, Matt Harvey, Head of Fixed Income Digital Sales and Client Execution Platforms, is working to improve the customer experience through increasing use of automation, digitisation and data-driven solutions. Harvey cites the enhanced data capture requirements of MiFID II, and the increased overheads this brought to the sales-trader workflow, as instrumental drivers towards greater digitisation of the fixed income desktop.

The firm’s Scout bot, built on the Symphony platform, enables salespeople to focus on the provision of higher-value conversations, information and services to clients. “The objective of the Scout bot is not to replace the human interaction between the client and the salesperson, but rather to augment it in a way that enables good data capture both for ML and for regulatory compliance requirements around how voice negotiations need to be handled for NatWest Markets and our clients, both now and as market structure continues to evolve,” says Harvey

He adds that “by presenting more data to our salespeople, we can also create a virtuous circle: you get people to engage more with patterns of client behaviour or particular aspects of the client and that creates better service.”

It’s not just traditional investment banks who see AI/ML as a means to enrich the customer experience and building stronger relationships. AiX is an AI-driven matching engine, built to act as a broker for institutional OTC traders in both cryptocurrencies and traditional asset classes. It automates pre- and post-trade data capture, handles multi-party negotiations, performs pre-trade credit checks, and passes trade details electronically to the settlement agency or clearing house.

The engine combines natural language processing (NLP), ML and cognitive reasoning technology to create a chat-bot, similar in concept to NatWest Market’s Scout, but with an additional execution capability.

 

Priyanka Jain, Business Development Executive at AiX, says that “AI/ML-based technology solutions are setting new standards for efficiency, accuracy, and regulatory best practice, not only in digital assets but also across all financial markets. I believe they will be welcomed by the front office, as this is their chance to re-skill and up-skill – to focus on more intelligent and high-value tasks like strategy, and leave the menial tasks to the technology.”

Further along the spectrum, there are firms focusing on using AI/ML to supplement human knowledge and to facilitate decision-making. For example, Liquidnet, an agency broker, acquired, OTAS, a trading analytics provider that uses customisable alerts highlighting opportunities, exceptions, and risks in global stock markets, delivered to clients in real-time and in plain English.

 

“So much work has been done already by the systematic and quant funds to really squeeze the lemon dry on data that impacts price. And that’s always been a speed race,” says Vicky Sanders, Global Head of Investment Analytics at Liquidnet. “We’re not in that race. We’re looking at how to best support the fundamental active investment community with data-led intelligence and insights. Investment decision makers typically apply a mosaic approach to investing and they will take many different data inputs, synthesise them and make their decision.”

She adds that “our focus for exception alerting relies on questions such as, are we targeting factors that might influence price? Are we targeting factors that might influence KPIs (key performance indicators) and fundamentals of a company?”

Forging ties

As to other solutions IPC, a communications and networking solutions firm for the financial markets community, has collaborated with GreenKey Technologies, a provider of voice software with integrated speech recognition to develop a range of cloud-deployed products based on AI/ML technology. The aim is to boost the sales-client experience and internal sales-trader workflow.

 

Their new workflow tools employ ML and NLP to convert a user’s voice quotes and trades into a streaming transcript as it unfolds in real time, which can then be deployed to populate blotters and order books. The technology also enables seamless integration of conversations across multiple channels including voice and chat. Other functionality parses quotes and trades alongside conversational raw text, creating a real-time internal price data feed. This allows firms to easily scan and determine the most current state of multiple conversations for faster trading against more up-to-date data.

“Even though it’s early days, ML advances are already allowing trading desks to realise faster execution, more efficient communications, and streamlined settlement and reporting processes right now, today,“ says Tim Carmody, Chief Technology Officer at IPC. “There are soft-dollar ROI (return on investment ) advantages with efficiency gains from real-time speech-to-text transcriptions, and capture of in-stream orders and quotes, and the ability to free up a workforce from repetitive dual-keying functions.

He adds that there is “also hard-dollar ROI in trading and compliance, where NLP can directly improve the bottom line through transaction cost analysis, cost-avoidance for fees and penalties, and surveillance in general.”

©BestExecution 2020