AI on the other end of the line (but you’re not talking to a computer)

Generative artificial intelligence (GenAI) is continuing to break new ground at a rapid pace. And with innovations in technology coming at us so fast, it’s easy to envision a future run by AI—self-driving cars, chatbots that write your work emails, and a virtual assistant that automatically buys your mother her favorite flowers on her birthday.

But as we’ve said before, we believe it’s the combination of people and AI, not AI alone, where the opportunity really lies. After all, is it really a thank you when your mom’s virtual assistant thanks your personal assistant for those flowers? And was it really even a gift in the first place?

OpenAI, the creators of ChatGPT, recently launched o1, a new family of large language models with enhanced reasoning capabilities that can tackle problems requiring multi-step reasoning and complex problem-solving strategies, opening the door for millions of new use cases for GenAI.

At Schwab we’ve been using AI models for many years to help drive scale and efficiency across our business, including machine learning models for fraud detection, and natural language processing models to support client services. And we’re thinking about advances in GenAI the way we think about everything: How can we use this to serve our clients better?

Since service is at the heart of who we are, we’ve focused on using GenAI to help our client-facing employees more easily, quickly, and accurately respond to complex service situations, enabling them to ultimately improve the client experience. We recently launched the Schwab Knowledge Assistant, which is a GenAI tool that provides a ChatGPT-like experience for our client service representatives. Using large language models, a type of AI that can recognize and generate text, the Schwab Knowledge Assistant searches Schwab’s internal knowledge management system, the Knowledge Center, and delivers answers to complex process and policy questions quickly and efficiently, including the sources for its answer. 

For a straightforward example to demonstrate how the tool works, let’s say a client called wanting to know if ETFs or mutual funds were more tax efficient. The rep could simply type, “Are ETFs or mutual funds more tax-efficient?” into the Schwab Knowledge Assistant. The tool would then search the Knowledge Center and pull together an easy to digest response in conversational language, along with its source material, giving the rep a simple and verifiable explanation of the potential tax benefits of ETFs over mutual funds to share with the client. 

Prior to the Schwab Knowledge Assistant, the client service professional would have searched for ETFs and then for mutual funds in the Knowledge Center and would have found a Knowledge Center article for both types of funds explaining the tax implications of each. The rep would then have had to read and comprehend both, then develop an answer for the client—a much longer process. The Schwab Knowledge Assistant now does this work for them.

We’re thinking of this tool as a “fast pass” to jump ahead in the knowledge line. Not as something to replace our experts. It can help our newest reps get up to speed faster and enable our more tenured reps to help clients with some of the most complex financial scenarios.

Ana Gonzalez, a Client Service and Support representative based in Phoenix, says the tool has been a game-changer for her. “The first day I used it, it felt like I had a supervisor right next to me, helping me answer questions,” she says.

And she’s seen the trickle-down effect for clients. “We are able to give clients answers faster and with more confidence. We have a higher success rate of resolving questions in one call, so they can get back to their lives and the other things they have to do,” she explains.

Early results show that the Knowledge Assistant may save up to two minutes for complex calls, which may not seem like a lot, but when you multiply that by about 7,000 frontline service professionals like Ana, serving thousands of clients each day, it can really add up, saving our reps and the clients they serve LOTS of time.

The Schwab Knowledge Assistant is an example of our belief that it will be the marriage of AI with the skills of our talented professionals that will ultimately deliver the best experience and outcomes for clients. We know AI will have a significant long-term impact on our business, especially when it comes to driving scale and efficiency and helping us serve our various client segments. We will continue to evaluate where this technology can add value for our clients and employees while keeping a thoughtful approach to ensuring we live up to our long-standing commitment to keeping clients’ data and finances secure.

Among the next use cases for AI that Schwab is prioritizing is the Schwab Research Assistant, which has been built on the learnings of the Schwab Knowledge Assistant. It provides a similar ChatGPT-like experience to our financial consultants and advisors, utilizing the Schwab Center for Financial Research (SCFR) as the underlying data source, which includes vast amounts of proprietary research and analysis, including expert perspectives about what’s happening in the market. One of the main objectives is to reduce task time required to research and summarize this data, allowing our teams to better serve clients with more easily accessible, timely financial planning information.

“Our plan is to continue to use the best technology to support our people, enabling them to have more meaningful conversations and deliver better, more personalized support to clients,” says Tim Heier, Schwab Chief Technology Officer.

“Our pride is our customer service,” says Ana reflecting on her role at Schwab. 

And she’s right. That’s why our focus with AI is on helping the people that help our clients. So the lesson here is that while soon it may become tempting and easy to rely on a virtual assistant for all things, maybe just use it to remind you of your mother’s birthday. And when it does, give her a call yourself, because you can’t replace human connection.