OpenAI’s ChatGPT has turbo-charged the rise of artificial intelligence. Individuals now have free access to world-leading generative AI models, and businesses are being forced to think about how to respond to these new capabilities. There is frenzied speculation about how this could change our lives and jobs. Tech titans such as Microsoft, Meta, and Google are moving quickly to integrate Large Language Models (LLMs) into their products, as are organizations in medicine, law, science, and more.
In this whitepaper, we present the facts, opportunities, and risks for asset managers, including a playbook for executives to navigate this new environment. Specifically, we:
- Examine the technical features of ChatGPT and models like it. We demystify the underlying technology, considering the model architecture and how they generate their outputs.
- Propose use cases and the possible benefits for an asset management organization. Removing barriers to powerful machine learning will commoditize simple tasks such as content retrieval and will free up (for example) investment and product teams to focus on market research and ideation. We look at use cases we are hearing about from our clients and assess LLMs and OpenAI’s GPT models’ features to identify where we see the most value.
- Consider the limitations of ChatGPT and other LLMs to highlight potential risks. ChatGPT is highly impressive, but it is not a panacea. LLMs such as ChatGPT have limitations in their current form, including data exposure, non-factual responses, unknowability, and bias. This can make them unsuitable for some use cases (such as regulated disclosures) without appropriate guardrails and may introduce business risk.
- Outline an ‘Executive Playbook’ for working with LLMs and OpenAI’s GPT models. We identify 7 considerations to help asset managers prepare for and navigate these new emerging AI capabilities and others like them. This includes how to assess models like GPT and their suitability for your business. We stress the importance of approaching your data science & analytics capabilities holistically and focusing on business needs, outlining a framework for evolving your business and AI/ML capabilities along with these new developments.
We believe LLMs like ChatGPT will ultimately be boosts for workplace productivity, particularly as they are integrated with existing enterprise applications like Office 365 and Salesforce. However, the models’ limitations make them less suitable for use cases such as generating proprietary insight for the investment process or automating regulated disclosures. Asset managers should explore and get ready for this new technology but do so with caution.
In our first whitepaper in February 2023, we set out Alpha’s framework for building a successful data science & analytics capability. The components of user experience, the insight ‘engine’, and strong data foundations all remain valid, and we emphasize these across this new paper. We recommend asset managers do the same, and we’d be delighted to discuss the approach, as well as our analytics technology solutions to help accelerate and guide your journey.
Contact Us
Alpha FMC has extensive experience working with asset managers to explore where and how data science and analytics can empower their business. For more information on how Alpha FMC can help your organization, please contact us here.
About the Authors
Natalie McIntyre
Director
Natalie is a Director, Head of Technology Services and a member of the Management Committee in North America. She has led large scale transformation programs and advised clients on business strategy, distribution management, investment operations and M&A due diligence for both retail and institutional asset managers. Most recently, Natalie has been on the forefront of leveraging Data Science to drive digital sales and generate investment alpha. Prior to joining Alpha, Natalie managed distribution operations and business intelligence at Bridgewater Associates.
Harry Taylor
Associate Director
Harry is an Associate Director and Co-Head of Data Science & Analytics in Alpha. Harry specializes in supporting asset managers drive efficient business growth using data within sales, marketing, and client service functions. Harry is also aligned to Alpha's Distribution Practice area which works more broadly with Alpha's clients in their digital and client facing capabilities.
Greg Beazley
Solution Architect
Greg is a Solution Architect based in Alpha’s London office. He is one of the lead architects in the Data Science & Analytics practice at Alpha and has extensive Python and SQL experience. Greg has been the development lead on several advanced analytics projects for asset managers, ranging from the creation of a custom data master to the development and deployment of data science models to assess client risk and optimize client prospecting.
Nigel Musiza
Solution Architect
Nigel is a Solution Architect based in Alpha's London office. He has 4+ years’ experience working with enterprise-scale businesses across Data Architecture, BI & Analytics and ML, with the majority spent at Alpha, focused on the Asset & Wealth Management industry. He has been Lead Developer on large-scale implementations of Data Mastering and Machine Learning capabilities, with particular focus on Trades/AUM data management, processing, and reporting.