Salesforce Redefines Wealth Management with AI

The AI Revolution

Of all the technological innovations since the birth of the internet, artificial intelligence (AI) is in a class by itself. By improving personalization and customer experiences, firms can future-proof and grow business over the long-term. And by prioritizing AI and a data-driven culture, wealth management firms can speed up their ability to analyze and act on vast amounts of data, accelerate productivity in the flow of work, and enable advisors, sales and service teams with AI-generated content, knowledge articles, and data-driven insights and actions to help them engage more efficiently and effectively.

Let your mind wander and say, how do we put generative AI to work in a meaningful way for the business?
If you spend some time understanding those concepts and how generative AI can improve experiences for your customers, then you can’t help but get super excited really fast.

Generative AI will fundamentally change the wealth management industry — and society — in a positive way. New roles will be created, such as generative AI security or generative AI customer service. Generative AI can make it easier to help customers manage their long-term financial planning, decide when to reconsider their investment allocations, or grow their retirement savings.

The immediate imperative is to learn and keep up with the pace with which generative AI technology is moving. Find the places in your organization where generative AI will produce the most benefit. For wealth management firms, acting now will offer the opportunity to start small, iterate and learn, and quickly benefit from the efficiencies generative AI drives. Start by identifying pain points in your customer and employee journey to understand where generative AI can produce the most benefit to your organization.

---

According to PwC’s Global Artificial Intelligence study, AI could contribute up to $15.7 trillion to the global economy by 2030, as well as up to a 26% boost in GDP for local economies. Or as Chris Hyzy, Chief Investment Officer for Merrill and Bank of America Private Bank, puts it, “AI is going to transform the global economy as surely as electricity and the steam engine did in their own times.”

The power and promise of generative AI is transforming every aspect of wealth management at an ever-increasing rate. The accelerated pace of change, driven by generative AI, is changing the way firms are designing next-generation experiences, powering sales teams, and approaching client engagement with intelligence and intuition.

Salesforce AI Terms and Definitions

Large Language Model (LLM): A type of AI algorithm that uses large data sets to understand, summarize, generate, and predict new content. LLMs can answer customer service questions, create personalized outreach, assist client research, and even generate code.

Generative Pretained Transformer (GPT): A neural network family that is trained to generate content. GPT models are trained on a large amount of text data, which lets them generate clear and relevant text based on user prompts or queries.

Hallucination: When a generative AI model produces new content that doesn’t correspond to reality or misinterprets its training data if the original data set was sparse, i.e., there were not many data points.

Retrieval-Augmented Generation (RAG): A layer that sits on top of the LLM and helps ground data. It takes a prompt and searches data (such as a Google search), finds the most relevant content, finds the most relevant paragraphs within data, and informs LLMs to build a very specific response

Explainability/Mechanistic Interpretability: Aspect of AI that ensures the system is doing the right thing (for example, following predetermined rules, such as adhering to established credit criteria).

Human in the Loop: Ensuring a human has oversight of a generative AI output and can give direct feedback to the model, in both the training and testing phases, and during active use of the system.

Ethical AI: AI that adheres to well-defined ethical guidelines regarding fundamental values, such as individual rights, privacy, nondiscrimination, and non-manipulation.

Salesforce AI for Wealth Management TODAY

By automating routine processes and providing actionable insights, AI liberates advisors to concentrate on high-impact activities. From generating personalized recommendations to optimizing portfolio strategies, AI streamlines operations, allowing advisors to deliver unparalleled value to clients.”

Although the concept of AI has been around since the 1950s, firms didn’t begin exploring AI for basic tasks like data management and customer service until the 1980s and 1990s. The explosion of big data and the refinement of machine learning led to more sophisticated AI applications. AI’s transformative role in wealth management extends to enhancing advisor productivity and enabling focus on value-added tasks.

As the industry embraces AI, advisors are poised to achieve new levels of efficiency and client satisfaction. According to John O’Connell, CEO and Founder of The Oasis Group, nearly 60% of advisors said in a recent study from SmartAsset that they are using or interested in testing out generative AI. Additionally, an Arizent study found that 87% of wealth managers believe that AI will be beneficial to the industry.

A CFP Board survey found that 31% of investors would be comfortable putting AI’s financial advice into practice without verifying it first. The importance of AI in wealth management is being increasingly recognized, offering unprecedented opportunities for efficiency, firm growth plans, and client service.

CDOs, CTOs, and CXOs are all trying to develop an AI strategy, test proof-of-concept AI use cases and embed AI and data deep into the flow of work. Data is the foundation to an amazing AI-driven experience — this is why many wealth firms are focused first on bringing together siloed data across their organizations.

They are then taking client data, transactional data and third-party data sources like CRM and financial planning data, as well as behavioral data, to get to a single source of truth,” she said. “Once that’s done, it’s all about leveraging that consolidated data to power LLMs, making prompt templates smarter, powering analytics, and generating alerts and other actions.

Predictive vs. Generative Salesforce AI

AI is the broad concept of having machines think and act like humans. But there are different types of AI, including predictive AI and generative AI, both of which help your teams work smarter and faster by automating routine tasks.

Predictive AI is designed to forecast outcomes based on historical data, and its primary goal is to predict future events or behaviors by analyzing patterns and trends.

Generative AI involves training a model to generate new data that is similar to the training data it was given. The output is typically some form of content — from text to images, to video, even computer code.

The Salesforce Generative AI Difference

Generative AI levels the playing field for large and small institutions. Capacity is no longer the limiting factor for the quality of your products. Consumers will ultimately reap the benefits from this kind of disruption.

While generative AI also relies on historical data, its primary goal is to learn the underlying patterns that are not restricted to replicating past examples. With generative AI, you can potentially create new financial strategies, create diverse customer engagement materials, or even assist in product design.

Salesforce AI Benefits and Use Cases for Wealth Management

With generative AI, both wealth management employees and clients are empowered with a powerful partner that drives better experiences all from natural language. Customers can receive hyper-personalized offers based on data they shared about their financial needs and goals, or resolve complex service issues through self-service. Advisors might leverage a virtual assistant to compose personalized emails or automate meeting summaries and related follow-up action items, just to name a few benefits.

When it comes to financial planning, AI can help augment advisors by keeping clients informed on their financial goal status and progress, analyzing portfolio performance and flagging risks to reaching client financial goals, or even suggesting when it’s time to update a financial plan.

A very exciting use case is to use chatbots to provide clients with self-service capabilities for routine tasks, such as requesting a money movement or resetting a password. The chatbot can validate the client quickly by comparing client answers with known information in the CRM. A client service associate can call the client to validate the request. This entire interaction reduces an enormous amount of friction from the daily activities of a wealth management firm.

Generative AI Benefits Across Lines of Business and Use Cases

Marketing

AI can create more personalized content, tailor messaging and offers, and automatically create web pages and campaigns, thereby increasing campaign effectiveness and ROI.

Example: Content creation. Generative AI can produce subject lines and email body copy that resemble human writing. By combining this with customer behavior predictions forecasted by predictive AI, advisors can send timely and relevant communications that match client needs to increase engagement.

Sales

AI can auto-generate client outreach emails, summarize interactions and next steps, and augment client and account research.

Example: Client insights. Generative AI can act like a personal data analyst assistant to use customer-shared data about their financial needs and goals to create hyperpersonalized offers. An AI virtual assistant can also develop client summaries, pull information together to automate meeting preparation and summaries, and compose personalized emails.

Service

AI can auto-generate service responses, write account and meeting summaries, and create knowledge articles to expand knowledge across any wealth management firm.

Example: Automated customer service. Generative AI can allow customers to resolve complex service issues, as well as routine tasks, through self-service. AI can also help advisors by analyzing clients’ portfolio performance and flagging potential risks.

Compliance and IT

AI can auto-generate document summaries, produce code from natural language prompts, provide chat-based assistance, and offer auto-completion for coding.

Example: Fraud detection and prevention. Generative AI can pick up on unusual trading behavior and contact a client to confirm the trade is legitimate.

Use cases

• Augment advisor productivity with financial planning updates.
• Act as advisor virtual assistant.
• Generate client summaries.
• Streamline and automate meeting prep.
• Generate next best actions. – Generate personalized content and knowledge articles.
• Self-service knowledge source for advisors and clients.

Generative AI Challenges

Wealth management firms should establish strict governance as well as a dedicated compliance framework for generative AI, ensuring that these systems are not just effective but also legally sound. Some challenges — and how to address them — include:

• Risk management: Firms should develop generative AI models in tandem with risk experts to tailor outputs to regulatory standards.
• Data privacy and security: Firms must enforce strict data governance policies, use advanced encryption methods, safeguard against using public LLMs, and continually update their security protocols to ensure that employees don’t share sensitive customer data.
• Bias and fairness: Regular auditing of generative AI models is essential to check for bias, risk of hallucinations, or language toxicity. Training these systems with diverse datasets and establishing guidelines to ensure decisions are fair and equitable are crucial steps.
• Change management: Integrating generative AI into existing infrastructure requires a strategic change management approach, encompassing clear communication, stakeholder engagement, and a phased implementation of generative AI models.
• AI expertise: Advisors and other users should invest in continual upskilling and training programs to build AI expertise across their workforce, while also tapping into new talent pools through collaborations with academic institutions and tech companies.
• Anticipation and adherence to new AI regulations: Firms are taking a cautious and conservative approach when it comes to AI use case development, awaiting any clarity on new AI regulation that may impact the way they design, build, and monitor AI data exchange. However, if they wait too long they risk falling behind their competition as they ride the AI wave.

Build a Trusted, Responsible, Generative AI Strategy with Salesforce

Responsible AI is AI with “good intentions” — and designing, developing, and deploying AI with good intentions is the best path forward for earning trust.

There is a trust gap when it comes to AI, preventing some customers from seizing the potential of generative AI. Many customers don’t believe AI is mature enough, yet safe and secure. They are worried about pending AI regulation, data privacy, data security, and accuracy, as well as concerned that an LLM will store and learn their sensitive information and company data. They’re worried about hallucinations and accuracy. Is AI telling us the truth? These are all real risks.

Five Key Guidelines for Developing Trusted Generative AI

1. Accuracy: Deliver verifiable results that balance accuracy, precision, and recall in the models by training models on your firm’s trusted, customer data. Communicate when there is uncertainty about the veracity of AI’s response, and enable users to validate these responses.
2. Safety: Make every effort to mitigate bias, toxicity, and harmful output by conducting bias, clarity, and robustness assessments. Protect the privacy of any personally identifying information present in the data used for training, and create guardrails to prevent additional harm.
3. Honesty: Respect data provenance and ensure you have consent to use the data. Wealth managers must also be transparent that AI has created content when it is autonomously delivered.
4. Empowerment: There are some cases where it is best to fully automate processes, but there are others where AI should play a supporting role to the human — or where human judgment is required. Identify the appropriate balance to “supercharge” human capabilities and involvement.
5. Sustainability: Develop right-sized models where possible to reduce your carbon footprint. When it comes to AI models, larger doesn’t always mean better: In some instances, smaller, better
   trained models outperform larger, more sparsely trained models.

Guidelines for Building Trust with Salesforce AI

Be responsible
Be accountable
Be transparent
Be empowering
Be inclusive

Implement Your Salesforce AI Strategy

Set clear objectives: Define what you want to achieve, and ensure alignment with your business goals and ethical standards. Be sure to develop metrics and thresholds for success, such as error rate and latency during pilot projects. Establish regular monitoring of AI systems to measure performance against objectives as well, and be prepared to iterate on your strategy based on feedback and measured outcomes.

Engage stakeholders: Involve them from the start, and bring in people from various departments, whether directly or indirectly involved. You want the most holistic view of AI’s potential impact as possible.

Assess technology needs and infrastructure: Identify the types of generative AI (like natural language processing, image generation, etc.) that are most relevant to your objectives. Determine if the current IT infrastructure can support generative AI’s computational demands or if upgrades are necessary.

Assess data readiness: Evaluate the quality and quantity of your data, and improve data collection and management processes if necessary. Ensure that data governance and ethical standards are in place to handle data responsibly

Data Security and Human Oversight

Many wealth managers are trying to find where artificial intelligence fits within their processes. I recommend that wealth managers consider AI applications in their middle and back office to streamline their operations team’s workflows first. This enables wealth management firms to start with well-defined use cases and well-defined processes to get their feet wet with AI. This also avoids many of the compliance questions raised with AI.

Despite the best safeguards against fraud or data leakage, the reality is that bad actors exist. Wealth managers need to be extra cautious and diligent to ensure quality and limit risk.

This includes:

• Conducting regular security audits and updates, and keeping software current to best protect against new threats.
• Reviewing relevant regulations regarding data privacy, intellectual property, and AI ethics.
• Developing or updating policies that reflect ethical AI usage, especially in the context of generative models that can create new content.

Also imperative is striking the right balance between automated processes and human oversight. Start with automating smaller, low risk tasks, then gradually scale up as the system proves its reliability and efficiency.

Vendor Selection

• Evaluate ethical standards: Choose vendors who share your commitment to ethical AI practices.
• Assess technical expertise: Ensure the vendor has the necessary technical expertise and experience in implementing AI responsibly.
• Consider long-term support: Look for vendors who offer robust support and maintenance post-implementation.

Finally… AI Implementation

• Take an iterative approach: Allow employees to experiment with generative AI on internal systems, so they can grow more comfortable and build their skill set.
• Continually monitor and evaluate: Regularly monitor generative AI systems — and their outputs — to ensure they operate as intended and adhere to ethical standards.
• Conduct employee training: Educate employees on how to interact with AI systems and understand their outputs