Written by Victor Slabodchikov and Paul Ananth
In our 2023 blog post, Asset Management Operations: What Can ChatGPT Do for Me?, we delved into the potential of ChatGPT as it stood then. From its inception, ChatGPT has impressed many with its ability to generate detailed responses across various domains. In our take last year, we concluded that while the technology holds immense promise, it is clear that ChatGPT cannot replace humans' nuanced intelligence and experience – at least not yet.
In this follow-up blog, we'll explore the evolving role of AI in asset management: security concerns, new enterprise features, trends and use cases, connecting the dots between ChatGPT's current state and the potential future landscapes of artificial intelligence in our industry.
A number of recent news articles (for example, Employees are banned from using ChatGPT at these companies) note that many large banks and asset managers are either considering or actively implementing bans on ChatGPT and similar generative AI applications in the workplace. The use of artificial intelligence in asset management has become a topic of concern for many organizations, given the potential risks to data security and privacy and the potential for damaging corporate reputation.
As we navigate the complexities of AI's impact on asset management, our journey continues with a closer look at ChatGPT Enterprise—a solution gaining traction this year that is designed to address the very challenges that have led to bans on its predecessor and limitations that narrowed adoption:
|
A word of caution: beware of AI hallucinations AI hallucinations occur when AI models generate incorrect information but present it as true. This happens because these AI tools lack the ability to think critically or consider whether what they're saying makes sense. The computer effectively gets carried away trying to give user an answer without really understanding the question or the answer itself. Even with some precautions, mistakes will still happen – AI remains prone to errors. |
Enterprise-grade Security: ChatGPT Enterprise is designed to keep private data private. It addresses safety concerns by building its institutional knowledge, while being aware of the difference between internal, confidential, and public data. It comes with layers of security to protect sensitive in-house information. It’s intended to be taught or configured to comply with the required privacy rules and procedures.
Enhanced Customization: Enterprise version offers new levels of personalization and setup. It allows organizations to customize the model to fit its environment. As a result, it can better understand and respond to questions related to the asset management industry and the specific subject matter at hand.
Ease of Integration: New version takes advantage of APIs, direct data sets access, and other integration solutions to work with the tools and systems businesses likely already use. Paired with a visualization solution, BI (Business Intelligence) or AI based, there is potential to simplify reporting, starting with internal operations and management data sets.
Team Collaboration and Knowledge Sharing: as we predicted last year, ChatGPT Enterprise dives into Knowledge Management and collaboration by providing a platform for accessing, analyzing, and sharing information among team members.
Data Analytics: ChatGPT Enterprise offers the potential to analyze large sets of data, in various file or data set formats, including structured and unstructured data from PDFs to Excel and CSV. ChatGPT can “understand” data, detects errors and anomalies, and derive insight that may not jump out or may take a while to compile for a human analyst. Specific queries can be asked in natural conversational language.
Presentations, Emails, or Document Composition: ChatGPT Enterprise, coupled with tools like Microsoft Copilot, could assist in compiling or editing content for various types of documents. Slides, bullet points, summaries, notes, minutes, charts, and other content can be created, reviewed, or transformed from one format to another by the AI. It can be used as inspirational early drafts or to proofread the working versions.
Document Review and Data Parsing: We often hear from clients the need to “read” and automate processing of unstructured data, such as PDFs and emails to extract, normalize, and store the contained information, transactions, instructions, etc. Many vendors have already started to implement AI or machine learning solutions, commonly referred to as Natural Language Processing (NLP) into their operations, albeit with human reviews. ChatGPT Enterprise offers the potential to simplify the adoption and implementation of these types of technology across various enterprise users and use cases.
As we wrote in our Outlook 2024 (Download Outlook 2024): Anyone who has experimented with ChatGPT firsthand can probably attest that this technology is still young and that 2024 is unlikely to be the year that investment managers implement NLP tools at scale. With that said, rooted in mass market technology, with the “move fast and break things” mindset, the pace of progress of AI solutions is staggering. The functional capabilities are continuously added. The safety and other gaps are being addressed.
Organizations should not dismiss tools like ChatGPT Enterprise and Microsoft Copilot outright. We would recommend budgeting for technical, business, procurement, legal, and regulatory reviews of these solutions in the coming years. Sooner or later, they will come to most enterprises either directly or through one of the software vendors or outsourcing partners.
ChatGPT Enterprise won’t write a Business Requirements document from an interview with a subject matter expert, yet, but it can make that email sound more or less formal, help convert a word document into a draft presentation, or put something on a page when you are just staring at a white screen and a blinking line.