The evolving role of AI in wealth management

Artificial intelligence-powered applications have quickly mushroomed within the wealth management industry.

Advisory firms can now choose from tools that can help them do everything from organizing and responding to client requests to hosting podcasts.

There’s plenty of excitement surrounding the potential these AI tools can offer. But financial advisors need some perspective on how they can use these tools to enhance client relationships while navigating potential pitfalls. 

The AI revolution in financial services

Large Language Models (LLMs) are the most popular type of AI systems and they have quickly penetrated our everyday lives, appearing in everything from document creation to email composition and spreadsheet analysis. 

You probably have already experimented with one of these tools. The current landscape features several leading players:

• OpenAI's ChatGPT: The pioneer that introduced LLMs to the mainstream and powers Microsoft's Co-pilot. GPT-4o now supports multi-modal inputs including text, audio, images, and video.
  
  
• Google's Gemini: Reaching a significant user base through integration with Google search, while also offering Gemma as an open model alternative.
  
  
• Anthropic's Claude: Distinguished by its focus on safety and alignment, Claude is designed to be helpful, honest, and harmless while minimizing risks like misinformation.
  
  
• Meta's Llama: An open model with a vast user base through Meta's platforms.
  
  
• DeepSeek: A recent entrant that made waves with claims of greater efficiency and reduced computational costs, impacting markets, particularly chip manufacturers.

Not all are created equally though. Each one has their own strengths and weaknesses. Some are more error-prone than others. Some offer more functionality in their paid versions than what’s freely available to the public. 

Here’s one thing to keep in mind when using any of these LLMs: They are constantly being “trained” on data, in other words, they ingest what users input and that becomes part of their “intelligence.” So don’t put anything in that would be considered sensitive, proprietary or your duty to protect, such as client information. Doing so would be exposing that data! 

Current applications

Financial institutions aren't merely experimenting with AI – they're actively implementing it across multiple functions. Here’s just a sampling:

1. Administrative Support: Meeting note transcription, client briefing preparation, and automated workflow creation.
   
   
2. Security Enhancement: Fraud detection through anomaly identification in transaction patterns.
   
   
3. Client Engagement: Chatbots providing immediate responses to client inquiries.
   
   
4. Personal AI Assistants: Tools like AdvisorEngine's JumpApp, which securely converts advisor-client conversations into tasks, notes, and compliance records, saving significant time for wealth managers.

Firms can benefit from enhanced efficiency these tools offer. Streamlined compliance documentation and an increased capacity to serve more clients are two emerging advantages to be gained from the newest crop of tools hitting the market. 

However, there are some areas that do not offer the robustness a fiduciary would require. 

There could be potential inaccuracies in financial analysis if models use outdated training data; there are real regulatory concerns around transparency and accountability of AI-generated advice, as well as the privacy risks when handling sensitive client information, which we already mentioned. 

There’s also a real risk that over-reliance on generative AI tools could also result in a cookie-cutter effect in client communications and outreach. The best approach still is to consider them complementary tools that augment human advisors' capabilities rather than replacing their judgment and interpersonal skills.

There’s no doubt automation will save advisors significant time and add meaningful value for their clients. However, we’re not the only ones accessing these tools.

Security concerns 

If there’s a real downside to the AI revolution, it’s that it brings substantial security risks. 

Today’s crude phishing scams and emails still manage to ensnare end clients and even RIA firms, costing them millions of dollars in losses. These same scams will become alarmingly sophisticated when powered by generative AI. 

Criminals are creating synthetic identification documents to fraudulently access accounts, while voice cloning technology can now replicate someone's voice from just a 15-second audio sample. In one particularly troubling case, a finance worker transferred $25 million after a video call with what turned out to be a deepfake "chief financial officer."

Financial institutions must remain vigilant as these technologies become more accessible and convincing. 

Beyond chatbots

We’re now witnessing the emergence of sophisticated AI assistants that go far beyond the basic chatbots you may have encountered while trying to buy an airline ticket or in online banking. 

The new AI assistants can access and analyze comprehensive client data – from investment portfolios and banking transactions to insurance policies and estate plans – to nearly instantly provide integrated financial insights that previously required extensive manual analysis.

These digital assistants will effectively function as financial copilots. They can continuously monitor client accounts, market conditions, and life events to proactively identify opportunities and risks that might otherwise go unnoticed.

By analyzing patterns in client data, AI tools can anticipate significant life events, often before clients themselves have begun considering them. This works for advisors, too: automating routine administrative tasks and data analysis helps uncover moments for meaningful outreach to clients, freeing advisors to focus on the emotional and high-judgment aspects of wealth management.

Agentic AI

Perhaps the most significant evolution in wealth management AI is the shift toward what experts call "agentic" systems – AI that doesn't just analyze information but can take appropriate actions based on that analysis.

Unlike earlier AI tools that primarily generated content or recommendations, these new systems have limited decision-making authority within clearly defined parameters. For example, an agentic AI might automatically rebalance a portfolio when market movements push allocations beyond predetermined thresholds, or trigger tax-loss harvesting opportunities when conditions are optimal.

Industry leaders emphasize these agentic capabilities are being implemented with extensive safeguards. Clients maintain ultimate control through preference settings, approval requirements for significant actions, and the ability to override automated decisions.

The human factor

When you have computers that can look and speak like humans, how does that impact actual people and relationships?

The growing prevalence of AI companions is occurring against a backdrop of increasing social isolation in American society. Harvard researchers found that 21% of U.S. adults are experiencing feelings of loneliness, highlighting a potential void that digital relationships could fill for many people.

According to Forbes, the six leading AI companion applications have collectively amassed approximately 52 million users, demonstrating substantial market demand for artificial relationships that simulate human connection. In fact, 25% of young adults believe artificial intelligence could eventually replace human romantic relationships, according to a report by The Institute for Family Studies. 

Already, AI is being positioned as a mental health resource. The Wall Street Journal has reported on the development and implementation of AI-powered wellbeing companions specifically designed for students.

These developments raise profound questions about the future of human connection across all relationship types, including professional ones such as the advisor-client dynamic in financial services. AI companions that can simulate empathy and understanding will challenge traditional assumptions about the need for humans in relationship-based services.

Be prepared

Strategically adopting AI tools will allow firms to do more at reduced costs. But the new tools come with their own unique challenges.

The need for heightened digital security measures will become even greater. Financial institutions must implement rigorous protocols governing system access and client data protection. The sensitive nature of financial information makes wealth management firms particularly attractive targets for cybercriminals, placing data security at the forefront of technology implementation considerations. Industry regulations are expected to continue tightening around AI systems that process client financial data.

Undoubtedly, AI will reshape the traditional advisor-client dynamic. The relationship model that has defined wealth management for generations is undergoing substantial transformation as AI is tasked with increasingly complex aspects of financial planning and portfolio management. This shift should prompt financial advisors to reconsider their core value proposition, and focus more on emotional intelligence, complex problem-solving, and life coaching aspects that remain distinctly human challenges.

AI will alter how firms operate, but the most successful advisory practices will be those that thoughtfully integrate technology while preserving and enhancing the human elements that clients ultimately value most.