AI agents that can work with brokerage accounts and place orders are sharpening a strategic question for retail brokers: whether AI strengthens their direct relationship with traders or hands more of that relationship to an outside provider. In October, Interactive Brokers introduced autonomous trading, allowing eligible clients to let connected AI submit supported orders within specified limits.
For broker operations, AI also offers a potential route to automate customer support, account opening, research and routine account questions. That could matter particularly for firms serving clients across languages and time zones. Any savings, however, need to cover model use, integration, supervision and compliance. AI-driven screeners, portfolio tools, earnings analysis and automated execution may simplify complex platforms, but adding an assistant alone may be easier for competitors to copy than joining it to proprietary data, execution, account records and risk controls.
The commercial case is not settled by higher engagement. More sessions, questions or trades do not by themselves show that clients stay longer, become more valuable or receive better outcomes. CFD/forex brokers could use AI for risk analysis, multilingual service, trading features and client retention, while facing the possibility that extra activity brings little durable customer value and raises conduct and compliance exposure. ETF platforms must weigh easier research and portfolio management against features that rivals can replicate; multi-asset firms could improve cross-market discovery and joined-up account analysis, but would also take on more complexity and expense.
Adoption outpaces material deployment
| Evidence | Figure or scope |
|---|---|
| Bank of England and FCA survey, October 2026 | 86% of responding financial services firms used AI |
| AI systems reported in the survey | 7% were classed as highly material to the businesses |
| Condor Capital Wealth Management digital advice report, Q2 2026 | Examined agents linked to funded brokerage accounts; the source gives no numerical result |
Who captures value across the chain
The survey covers financial services as a whole, not retail brokers in particular, so its figures show a gap between adoption and business significance rather than proving that brokers are making money from AI. More useful evidence for an individual firm would come from its own performance data. Even then, operators would need to account for the possibility that people who select AI tools were already more active or technically confident.
Brokers can earn through trading charges, spreads, assets held on their platforms and lower service costs. Execution, integrated accounts and risk management may help them stand apart. But if third-party assistants become the place where investors compare providers, research markets and initiate trades, brokers could lose control of discovery and face sharper competition on price and execution. AI developers, meanwhile, may charge for API use, subscriptions or enterprise deals and gain influence by controlling the interface. Their returns could be limited by inference expense, rival models and reliance on brokerage APIs. Investors may save time, review more instruments and automate routine portfolio work, although charges, trading costs, poor decisions and overtrading can erase those gains.
The UK FCA raised the interface issue in its July 2026 Mills Review. It warned that AI agents could shift influence towards firms that manage customer-facing tools, and that product comparisons generated by AI might weaken the effect of established brand loyalty. Some firms could benefit if agents select their products, while others may have to appeal to systems ranking options by price, value, risk or suitability.
A Q2 2026 digital advice report from Condor Capital Wealth Management tested AI agents connected to funded brokerage accounts. It described capable tools alongside significant setup obstacles and open concerns about reliability, security and liability. The researchers cautioned that brief tests could not establish how the tools would perform in future. The account of the competitive debate was first reported by Finance Magnates.
The central infrastructure question is whether brokers can retain enough of the client relationship to justify what they spend, or whether an external agent becomes the main route into the account. Better analysis does not guarantee better investor returns, and broad AI adoption does not establish that deployment is materially changing broker economics.
What this changes for operators
Broker technology teams should map which parts of the stack an AI feature touches: account data, permissions, order submission, execution controls, customer support and audit records. Before connecting an agent to a funded account, operators need clear answers on the orders it can send, the limits it must respect and how activity is monitored. The source does not specify the technical controls used by the cited platforms.
For each deployment, operators should test whether service savings outweigh model, integration, monitoring and compliance costs, and measure retention or service outcomes against a suitable comparison group. They should also ask who owns the customer interface, how agent access is revoked, and how account activity is reconciled and reviewed. Those answers will show whether AI is an operating improvement for the broker or mainly a new route for another firm to reach its clients.
Platforms · Software · Providers · Markets