Where Does AI Fit in Your Firm’s Future?

The demand for AI

The demand for AI in wealth management was clearly outlined in the Morgan Stanley Wealth Management Pulse Survey from May 2023. The survey noted 63% of investors would be interested in working with a financial advisor who leverages AI. The number was overwhelmingly high among younger investors ages 35-44 where 85% were interested in working with an advisor who leverages AI. The belief that AI will not replace the advisor-client relationship was equally high at 84% believing that the relationship is critical.

A brief history of AI

AI has been a key focus of research for more than six decades, with the earliest roots in the 1950s. In 1956, a historic meeting at Dartmouth College brought together some of the greatest minds in computer science and philosophy to discuss ways that machines can acquire human intelligence. The American cognitive scientist John McCarthy coined the term "artificial intelligence" during this meeting, and research into AI began to take off shortly thereafter.

One of the earliest and most successful AI applications was IBM's Deep Blue, an expert system of AI, which defeated world chess champion Garry Kasparov in 1997. Since then, AI has been applied to a wide range of domains including robotics, natural language processing, image recognition, medical diagnosis, and autonomous driving. Developments in machine learning have allowed computers to outperform humans on tasks such as identifying objects in images or recognizing spoken words.

Applications of AI

Natural language processing

ChatGPT is a large natural language processing (NLP) tool and an application of AI. The technology behind large language models like ChatGPT is like the predictive text feature you see when you compose a message on your phone. Your phone will evaluate what has been typed and calculate probabilities of what’s most likely to follow, based on its model and what it has observed from your past behavior.

Unlike the phone’s predictive text (PT) feature, ChatGPT is generative (the G in GPT). It isn’t making one-off predictions; instead, it’s meant to create text strings that make sense across multiple sentences and paragraphs. The output is meant to look as though a person wrote it, and it should match up with the prompt.

Generative predictive text is an application of AI designed to have human-like conversations and much more with the chatbot. The language model can answer questions and assist with tasks, such as composing emails, blog posts, essays, and writing computer code. The language model can also analyze writing that you provide to determine your writing style, also known as your “voice.” Generative predictive text can then design its output to mimic the writing style of the author who provides the GPT model with their input.

Many wealth management firms like Orion, Broadridge, and FMG have already launched ChatGPT integrations. FMG leverages its existing content library and uses ChatGPT extensions to write the text for the advisor to share the third-party content on social media with a caption that is unique to each advisor who uses the feature. Future iterations may use the feature in many GPT solutions where the AI application can re-write an existing article. This feature is available in tools like Jasper.ai that also include integrations with plagiarism checkers, such as Copyscape.

Artificial neural networks

Artificial neural networks (ANNs), another type of AI application, were used to effectively predict stock prices. In 1994, David Lowe published a paper, “Novel Exploitation of Neural Network Methods in Financial Markets,” which discussed how AI could improve predictions in financial markets. The paper examined prices from seven component stocks to predict future FTSE-100 Index price. Chi Hau Chen published a paper in the same year using four daily market prices – opening, high, low and closing – to predict future market prices. Lowe and Chen demonstrated successful application of neural networks in predicting stock market prices. This research has been highly influential in the financial sector and has helped to spur further advancements in AI-driven trading algorithms.

Machine learning

Machine learning uses algorithms to discover patterns and generate insights from large amounts of data. You’ve been exposed to machine learning for years with recommendations from any number of vendors including Amazon, Netflix, and Starbucks. You’ve been exposed to machine learning when you fly and the Transportation Security Administration uses facial recognition when you check in or if you’ve ever been to a Disney theme park. Platforms from Facebook to Instagram and Twitter are using big data and AI to enhance their functionality and strengthen the user experience.

Machine learning has excellent wealth management use cases. My firm, the Oasis Group, published a white paper titled Next Best Action Technologies Make One-To-One Engagement Possible that examined the use of machine learning to generate next best actions that financial advisors can use in meetings with their clients.

Defining your AI project

Firms are struggling to define attainable AI goals. The Accenture study found that:

• 50% of respondents stated their wealth management companies face difficulties implementing their AI vision.

• 55% said their companies' AI tools and insights are difficult to use.

• 64% said their company is launching too many AI pilots at once to adopt the technology.

Any wealth management firm seeking to leverage AI in their business (which should be every wealth management firm) should follow some basic steps to ensure that they can clearly define their AI goals and the projects to reach them.

1. Establish a center of excellence

Implementing a successful AI strategy requires a diverse team to bring the necessary technical and business skills to the project. Teams should include AI specialists, business leads, and IT leads. Look outside of your organization for AI specialists if you do not have the expertise in your organization.

2. Identify business opportunities and set priorities

Your center of excellence should identify the business processes where AI can add the most value, document those processes if they are not already and determine the potential returns that AI can deliver. Once you’ve identified the business processes that can benefit the most, then develop your use cases and requirements. A set of requirements helps your AI project to have a clear set of goals that it needs to achieve to be successful.

3. Select and commit to a limited number of projects

Executives must select an attainable list of promising AI projects and commit to delivering minimally viable products (MVPs) that satisfy the requirements. Proof of concepts and pilots will not demonstrate success with AI. AI needs to be trained to be most effective. Commit to a small number of projects and deliver the MVPs.

4. Identify the AI application best suited to achieve your goals

Once you’ve identified the MVPs, identify the application of AI that best suits each project. Leveraging old content within the firm with an AI refresh may best be accomplished with a GPT application. The next best action will be best accomplished with a machine learning application. Review your projects and then select the applications that are best suited to deliver your MVP.

5. Own your data

AI requires a massive amount of good quality data to generate the best results. Executives must have a data strategy to gather, clean, move, and store all that data and deliver it to the AI systems at the right time and speed. Create the role of a data steward within the organization who is responsible for collecting, collating, and evaluating issues and problems with data. This key role will define the data flows within the firm and establish rules of how to resolve data issues. Firms need a strategy to own their data and should have conversations with their business partners to establish clear understandings of data ownership.

6. Address security, privacy, regulations, legalities, and ethics

AI comes with significant security, privacy, regulatory, and compliance concerns as well as legal issues and ethical implications. Address these areas from the start and as AI programs mature. Your center of excellence and leadership team should ask the questions “How are we protecting this data?” and “What if this goes wrong?"

History has innumerable examples of companies that did not adapt fast enough to changing market conditions and who are now out of business. AI will change the wealth management profession, and the pace of this change will be dizzying. Firms without a clear AI strategy or the ability to execute it will be at a significant disadvantage. It is incumbent upon today’s leadership teams to evaluate their data and initiate AI projects now.

John O’Connell, founder and chief executive officer for The Oasis Group, specializes in helping wealth management and technology firms to solve their most complex challenges. His newest online training courses serve as a leading source of education for financial professionals at all levels in their careers. With modules ranging from cybersecurity to custodian markets and more, https://training.theoasisgrp.com/ enables firms and enterprises to upskill, learn at their own pace and rewatch lessons to reinforce specific learning objectives. Get an additional 20% off any course with coupon code ADVISORPERSPECTIVES.

References

CHEN, C. H. (1994). NEURAL NETWORKS FOR FINANCIAL MARKET PREDICTION. PROCEEDINGS OF THE IEEE 1994 INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE, 2, 1199-1202.

LOWE, D. (1994). NOVEL EXPLOITATION OF NEURAL NETWORK METHODS IN FINANCIAL MARKETS. PROCEEDINGS OF THE IEEE 1994 INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE, 6, 3623-3628.

AI IN WEALTH MANAGEMENT A FINANCIAL ADVISOR STUDY, ACCENTURE, JUNE 22, 2022.

HTTPS://WWW.WEALTHMANAGEMENT.COM/TECHNOLOGY/WEALTHTECH-FIRMS-AND-ADVISORS-AI-HAS-ENTERED-CHAT

---

A message from Advisor Perspectives and VettaFi: To learn more about this and other topics, check out our most recent white papers.

More Large Cap Growth Topics >