GET digital brain

Image: Getty Images

ChatGPT is riding a hype cycle, but financial services have serious plans for natural language processing. Liz Lumley reports. 

It has been hard to avoid discussions around the launch of ChatGPT over the past few months. The buzzy service is an artificial intelligence (AI) chatbot developed by OpenAI built on top of OpenAI’s GPT-3 family of large language models and has been fine-tuned using both supervised and reinforcement learning techniques. Despite the hype, the possibilities offered by large language models have many in financial services planning strategically. 

Machine learning (ML) and AI in financial services have often been trained on quantitative data, such as historical stock prices. However, natural language processing (NLP), including the large language models used with ChatGPT, teaches computers to read and derive meaning from language. This means it can allow financial documents — such as the annual 10-k financial performance reports required by the Securities and Exchange Commission — to be used to predict stock movements. These reports are often dense and difficult for humans to comb through to gain sentiment analysis. By using NLP, investors can quickly analyse the tone of a report and use the data for investment decisions. In addition, NLP models can be used to gain insights from a range of unstructured data, such as social media posts.

According to Fortune Business Insights, the global market size for NLP could reach $161.81bn by 2029. 

As part of this AI trend, Symphony, which provides an infrastructure and technology platform for financial markets, has acquired Amenity Analytics, an NLP data analytics system aimed at portfolio managers, research professionals, analysts and other financial markets participants. 

The system extracts data insights through research quality assurance, tagging and key drivers from a variety of content types, including earnings call transcripts, news, social media, filings and research, among other publicly available sources. 

“Amenity has developed impactful use cases that tackle real time environmental, social and governance insights, targeted content delivery and information overload — all key to the future of the finance world — and now they’ll be available to the over 1000 institutions Symphony serves,” says Nathaniel Storch, Amenity Analytics CEO.  

Mike Lynch, chief product officer at Symphony, says the Amenity service has three main use cases. First for a new asset manager or hedge fund. “We're enabling them to generate alpha because they’re finding signals in noise that they would not be able to find without our models,” he adds. “We’ve [tested] how the model performs in terms of extracting insights from earnings and news that enables them to make better investment decisions.”

The second use case is for firms that cover thousands of companies — every company above a certain market capitalisation — which does not have the ability to read every earnings transcript end to end, says Mr Lynch. The third use case concerns better understanding the customer and improving customer engagement. 

“We’re able to help them refine which customers they are servicing, how they’re servicing them, and where they need to make their products better,” he adds. 

Symphony used a range of both open-source and proprietary capability to deliver the NLP product and ensure the insights are meaningful via back testing. Mr Lynch says investment and trading businesses have some of the lowest tolerance for mistakes and errors.  

“We can’t have an 80% accuracy rate ... that causes real problems from all sorts of ways,” he says. “The models are great for accelerating workflow and getting answers faster, but they’re not worth anything ... if the answer isn’t right.”

The use of AI and machine learning is fast becoming more impactful and influential in financial services. According to a 2022 survey from the Bank of England (BoE), 72% of UK financial services firms use ML and that percentage is expected to grow by three and a half times over the next three years. 

Around 80% of respondents to the BoE survey have data governance frameworks in place, with model risk management and operational risk frameworks also commonplace. However, most did not view the use of ML as high risk, with top risks including data bias and representativeness, as well as the lack of explainability and interpretability of ML applications.

PLEASE ENTER YOUR DETAILS TO WATCH THIS VIDEO

All fields are mandatory

The Banker is a service from the Financial Times. The Financial Times Ltd takes your privacy seriously.

Choose how you want us to contact you.

Invites and Offers from The Banker

Receive exclusive personalised event invitations, carefully curated offers and promotions from The Banker



For more information about how we use your data, please refer to our privacy and cookie policies.

Terms and conditions

Join our community

The Banker on Twitter