Schwab CEO Says Investors Must Set Rules for AI Trading Agents

Charles Schwab is adding artificial intelligence tools for trade planning while restricting how outside agents reach customer accounts, as CEO Rick Wurster says clients must define the limits for automated trading.

At Schwab’s Fall Business Update on Oct. 15 (ET; UTC date: Oct. 15), Wurster said the company was unlikely to provide its own algorithms or trading content for client strategies. “I think it’s a fool’s errand and could lead to a lot of trouble,” he said. Schwab is rolling out an AI capability to help clients consider strategies and make a plan.

Wurster said client trading through large language models (LLMs) and application programming interfaces (APIs) had doubled in about six months as of Oct. 15. In an Oct. 16 interview (ET; UTC date: Oct. 16), he said an AI agent should trade for a client only under rules set by the investor. “An AI agent should only trade on your behalf if you give them the rules of the road,” he said. His examples included researching a stock and placing a limit order.

Schwab blocks unauthorized agents that try to sign in as clients. The firm plans to use Model Context Protocol (MCP) servers to connect external AI models with account information for balance checks and research, while preventing those models from moving money. Wurster described the number of blocked attempts as in the “10s and 20s” per week.

He also raised questions about responsibility when an outside agent acts without clear control: “We can’t control it, the client can’t control it, something could go wrong and who’s liable?”

Schwab introduced Charley, a conversational AI assistant that gives U.S. retail clients access to Schwab tools and insights, and partnered with Anthropic on Claude for Financial Advisors.

The brokerage averaged 10.7 million trades per day in Q3 2026, up 44% from a year earlier. The quarter ended Sept. 30, and the figure covers all trading, not only activity involving AI.