Webull working with regulators as AI trading evolves

Webull is allowing clients to connect AI assistants including ChatGPT, Claude and Grok to brokerage accounts to retrieve data and, for supported products, place, modify or cancel trades through natural-language commands. The broker is actively engaging regulators over whether an AI agent is merely executing client instructions or making investment decisions, while retaining investor responsibility for trades. Webull has implemented controls including order previews, position-size caps, symbol restrictions and read-only access, signaling a measured rollout of agentic trading.

Analysis

The near-term economic value for BULL is not autonomous trading volume but lower customer-service friction, higher engagement, and potential conversion of less sophisticated users into funded, active accounts. If natural-language workflows reduce abandonment during onboarding or order entry, the upside accrues through net new assets, margin balances, options activity, and securities-lending inventory; however, these benefits are unlikely to be material before several quarterly reporting cycles. HOOD is the closest competitive read-through because its valuation embeds consumer-product velocity, while IBKR and SCHW are better positioned to monetize AI as a service and adviser productivity tool rather than as a retail trading interface.

The binding constraint is regulatory classification, not model capability. A finding that a broker or integrated agent is providing recommendations, exercising discretion, or inadequately supervising prompts could require suitability controls, recordkeeping, disclosures, and potentially human-review processes—turning a low-cost feature into a compliance-heavy one. The asymmetric downside is an AI-induced erroneous-order event or account-takeover incident: even limited direct losses could prompt regulators to slow rollout industrywide and compress retail-broker multiples on higher expected compliance and litigation expense.

Consensus may overestimate the immediate threat that AI interfaces disintermediate brokers. Custody, best execution, clearing, market-data entitlements, identity verification, and account controls remain broker-controlled bottlenecks; the likely structural outcome is that brokers become infrastructure providers to agents rather than being bypassed by them. The more important 6-18 month question is whether AI makes trading sufficiently easier to raise low-quality turnover, which would lift transaction-linked revenue in benign markets but increase regulatory and credit-loss cyclicality during volatility spikes.

Key Decisions for Investors

• Maintain BULL as a watch-list long rather than initiating on this development alone; require evidence of sequential funded-account growth, net new assets, or a disclosed AI-driven engagement metric over the next 1-3 earnings reports. Falsifier: added compliance expense without measurable customer-acquisition or asset-growth acceleration.
• Prefer a 6-12 month relative-value basket long IBKR / short HOOD if AI-agent adoption becomes a sector narrative: IBKR has more diversified global and professional-client economics, while HOOD has greater exposure to retail-risk sentiment and transaction-intensity normalization. Cover if HOOD demonstrates materially faster net deposits and subscription monetization than IBKR.
• Monitor SEC/FINRA guidance, enforcement actions, and broker disclosures on agent-initiated orders as the primary catalyst calendar. Any requirement for agent-level suitability, supervisory review, or expanded record retention is a negative read-through for BULL and HOOD and favors larger compliance platforms such as SCHW and IBKR.
• Do not underwrite incremental retail-broker trading revenue until order-flow, options-volume, and margin-balance data confirm behavior change; a market-volatility-led volume spike would be an unreliable signal because it may raise customer-credit and reputational risk simultaneously.