Charles Schwab announced on September 30 the launch of Charley, a new generative AI-powered assistant designed to give retail clients a more conversational way to access information, tools and services. Beginning in October, Charley is rolling out in phases to eligible U.S. retail clients on Schwab.com and Schwab Mobile. Through voice or text, clients can ask questions, receive suggested next steps, complete select tasks and connect with a Schwab representative. In an interview with Traders Magazine, Amy Ruegg, Head of Digital Retail at Charles Schwab, discusses what the firm learned from testing Charley, how it is addressing accuracy and client data, and where AI fits alongside human expertise.

What prompted Schwab to launch Charley now, and what did you learn from the employee pilot?

Schwab has been steadily expanding how we use AI to make investing more accessible, and Charley is the next step in that evolution. Ahead of the phased client rollout, thousands of employees helped test and refine Charley, helping us strengthen its reliability, accuracy and ease of use while learning how people naturally interact with a conversational experience.

How is Charley different from a traditional chatbot or Schwab’s existing digital tools?

Charley goes beyond a traditional chatbot by helping clients move from questions to actions faster through a more natural, conversational experience. Through voice or text, clients can ask questions in their own words, receive clear information and suggested next steps, complete select tasks, and connect with a Schwab representative when needed. Charley can also take certain eligible actions on a client’s behalf, rather than simply explaining how to complete them.

Where do you draw the line between tasks Charley can complete and decisions that require human involvement?

Schwab research shows clients are interested in AI playing a role in their investing experience, particularly when it is paired with access to people. In a recent survey of retail investors, 65% said AI can either help execute routine tasks or play a meaningful role in investing when paired with human expertise. Clients are most comfortable using AI to answer questions, find account information, and handle routine updates, while preferring human involvement for more meaningful decisions and actions.

That balance is central to how we’re building Charley. Charley acts only at the client’s direction and requires confirmation before making a change. And when a client wants or needs human support, Charley can help connect them with Schwab. AI can make our clients’ financial lives easier, but trust and access to people remain fundamental to the Schwab experience.

How are you addressing accuracy and the risk of incorrect or misleading AI-generated information in an investing context?

Charley combines the power of large language models with Schwab’s data, expertise, security and safeguards. Schwab uses testing, safety guardrails and human oversight to evaluate Charley’s accuracy and performance. Charley also provides links to relevant Schwab resources so clients can review source information and access additional context when needed.

How do you protect client data and privacy when Charley uses large language models?

Charley operates within Schwab’s authenticated digital environment and accesses only the information needed to respond to a client’s request. Schwab has also taken steps to ensure client data is not used to train Charley or the underlying AI models.

How do you expect Charley’s capabilities to evolve, particularly around personalization and investment-related tasks?

We’re rolling out Charley in phases, with the goal of making it increasingly useful and relevant to each client over time. That means moving beyond answering questions to using client context to help people understand what matters, explore their options and take appropriate actions at their direction. Ultimately, we see AI as another way to deliver on Schwab’s role as a trusted partner, bringing together digital intelligence, Schwab’s expertise and access to people when clients need them.