Anthony Denier can build a three-month Apple chart with moving averages and an S&P 500 overlay on Webull in about 30 seconds. He figures a new customer could spend 15 minutes hunting down the same tools.
“If I wanted to pull up a three-month Apple chart on my Webull app, and I wanted to put in some moving averages, maybe a broad-scale S&P index overlay, I can do that very easily because I’ve been using Webull since we started,” Denier, Webull’s group president and U.S. CEO, said in an interview with TBBW Magazine. “I know all the locations of all the different analytics, and I can probably do that in 30 seconds. A new user would probably take 15 minutes to figure out how to do all those things.”
On Aug. 4, Webull began letting customers connect their brokerage accounts to ChatGPT, Claude and Grok. The outside AI assistants can access account information, pull market data and, for supported products, place, modify or cancel trades. Customers can require a preview before an order goes through, cap its size or value, restrict trading to selected symbols or make the connection read-only. They authorize access through Webull rather than giving brokerage credentials to ChatGPT, Claude or Grok. The connectors followed Webull’s Model Context Protocol server, introduced less than two months earlier, which lets investors use natural-language prompts to access market data, account information and trading functions.
“AI is going to redefine the way that someone interacts with their brokerage account,” Denier said. “If you think back in the ’80s, you had to make a phone call, and then in the mid-’90s, you could log in for the first time. In the teens, you’re actually trading on your smartphone. The next generation is all going to be natural language.”
“The next generation that we’re going to really embrace is going to be just hitting the microphone button on the Webull app and saying, show me a three-month chart of Apple, overlay 30-day, 50-day, 200-day moving averages, put on top of the S&P, and instantaneously you’re looking at that chart,” he said. “No more clicking of buttons, no more figuring out, like, where’s the dropdown of that?”
An employee who wants a suite at a Rays game for clients can now use natural-language AI to bypass some of the bureaucracy that once sent the request from manager to manager before it reached Webull’s chief financial officer. “You [used to] put in a request that gets approved by manager, manager, manager, CFO,” Denier said. “Now it’s all natural language.”
That automation hasn’t cut Webull’s payroll so far. The company has 277 employees in St. Petersburg and expects to surpass 300 within about six months. “I haven’t seen any reduction of headcount,” Denier said. “I don’t foresee any reduction of headcount because of AI. That’s not saying there won’t be in the future, but we are a growing company.” He added: “We’ll probably easily break 300 in the next six months. We’re continuing to hire aggressively.”
Denier is requiring every member of Webull’s U.S. staff to complete an Anthropic AI course by year-end, covering basic use of the tools and what company information employees should and shouldn’t enter into them. He took the course himself. “Everyone needs to be certified, and it’s just bare-bones basics, kind of what to do, what not to do, especially when it comes to information,” Denier said. “That’s how I learned this.”
Webull lets an AI agent act on an investor’s instructions, but the stakes change once those instructions can produce a trade. The software cannot currently decide when to buy or sell on its own. Customers can limit the agent to selected symbols, cap order size or value, require a preview before submission or switch the connection to read-only access. Webull also warns that AI agents can misunderstand instructions or act on delayed information.
Webull has not said what it will eventually allow an AI agent to do without a fresh customer instruction. Denier said some of those uses could raise regulatory questions over whether the agent is carrying out a customer’s decision or making an investment decision itself. “We work directly with regulators,” he said. “We have open dialogue almost every single day, talking about, ‘This is what we’re looking to do. What would be an interpretation? How would you look at this?’”
“We need to make sure we use education and small steps toward utilizing this new technology,” Denier said. “This is what we’re rolling out. This is why we see the benefit. These are the protections we put in place.”
Webull didn’t disclose its own options volume, but Denier pointed to zero-day options, contracts that expire the day they are traded. “The adoption of zero-day options trading is just huge,” he said. “We’ve never seen the amount of options I see going through our clients.”
Average daily U.S. listed-options volume reached 72.8 million contracts in the second quarter, up more than 19% from a year earlier, according to Cboe. Zero-day options volume rose 46.2% year-to-date to more than 20 million contracts a day.
“Everything has to be customer-initiated AI today,” Denier said. “In the future that will change.”