Traders’ Toolbox: AI Order Placement Is Here

Artificial intelligence has become a popular tool for research, market monitoring, and idea generation. But what happens when AI moves beyond analysis and begins placing trades? Interactive Brokers’ Autonomous Trading framework is designed to explore that possibility while maintaining strict controls over what an AI can do and how much capital it can access.

At the center of the experience is the AI Partition, a separate trading environment that can be connected to an AI application. Rather than granting an AI access to an entire brokerage account, clients allocate a limited amount of capital to a dedicated partition where the AI can operate within a defined set of rules.

The current functionality focuses on a straightforward set of products and order types. AI may trade stocks, ETFs, and listed equity, ETF, and index options. Supported order types include market, limit, stop, and stop-limit orders. Time-in-force instructions are likewise restricted to Day, Overnight, and Overnight + Day orders.

What won’t the AI be trading? Several higher-risk or more complex activities have been deliberately excluded. Penny stocks, complex leveraged products, fractional shares, short sales, and options strategies requiring margin borrowing are all off limits.

Beyond trading capabilities, AI agents can access familiar account features such as market data, watchlists, alerts, holdings, orders, and trade history. This allows the AI to evaluate positions and market conditions before taking action.

One of the more interesting aspects of the platform is its flexibility. Any large language model that supports the Model Context Protocol (MCP) can connect to IBKR’s public MCP server. Whether you’re using Claude Desktop, Claude Code, ChatGPT, or another compatible AI platform, the connection process is designed to be straightforward. In many cases, users simply provide the MCP endpoint to the AI platform and establish the connection.

Of course, autonomous trading raises an obvious question: What safeguards are in place?

The answer begins with client control. At any time, clients can disconnect their AI through a dedicated “kill switch” available within Client Portal. If an AI begins behaving in a way that makes you uncomfortable, the connection can be terminated immediately.

Funding limits provide another layer of protection. Clients can allocate up to 10% of their account equity to AI Partitions, subject to a lifetime minimum funding amount of $1,000 and a maximum of $25,000 across all AI Partitions. For example, a $2,000 account may fund an AI Partition with $1,000, while a multimillion-dollar account remains capped at $25,000.

Looking ahead, Interactive Brokers plans to introduce additional user-defined guardrails. Future controls are expected to include limits on order frequency, maximum order sizes, and minimum account value thresholds, giving clients even greater oversight of autonomous activity.

The idea behind Autonomous Trading isn’t to remove the investor from the process. Rather, it’s to provide a controlled environment where traders can experiment with AI-driven decision making while maintaining clear boundaries around risk, capital, and execution. As AI technology continues to evolve, those guardrails may prove just as important as the intelligence behind the trades themselves.

1.  To qualify for Autonomous Trading, clients must first use Review & Approve mode, in which they manually approve each AI-generated order before submission. After completing at least 10 trades, each involving securities worth $100 or more, over a two-week period, clients with eligible account types and entities can enable Autonomous Trading. 
2. Soon we will have an eligbility test users could take to qualify in lieu of the trading requirement. 

 Those interested in learning more can visit: https://www.interactivebrokers.com/en/trading/ai-integrations.php?menu=B