How is AI changing brokerage — and where does it actually change the business?
Updated October 2026
The label "AI-native brokerage" can hide very different developments. Much of what is live today improves research, service, operations or advisor productivity without changing the underlying business.
More consequential possibilities begin where AI changes the customer proposition, the economics of serving investors or the distribution role of the broker itself.
Advisor copilots, research retrieval, customer-service augmentation, KYC and document processing, compliance automation and post-trade exception handling.
These applications can improve productivity, service or operating economics while leaving the underlying brokerage proposition largely unchanged.
AI-supported investment discovery, portfolio explanation, tax and account intelligence, conversational interaction and more personalised investment propositions.
Here AI becomes more visible to the investor and may change how customers interact with, understand and use their brokerage relationship.
New distribution models in which third-party AI interfaces connect to regulated brokerage infrastructure, or propositions that make previously uneconomic customer segments viable.
This is where the more structural questions begin: whether AI can change who owns the customer interface, how brokerage capabilities are distributed, which customers can be served economically and where value accrues.
The most visible customer-facing signals today are emerging in retail and self-directed investing — across different stages of the shift from AI inside the proposition towards AI potentially changing distribution.
Lightyear AI brings AI-generated market intelligence directly into the brokerage experience: explaining price and portfolio movements, summarising relevant market developments, supporting investment discovery and presenting arguments for and against individual investments. AI changes how investors discover and understand information, while the customer relationship and investment interface remain within Lightyear.
Revolut's AIR brings conversational interaction into the customer's financial relationship, including investments. Users can ask about stock prices, portfolio performance and investment information alongside other financial questions and tasks. This changes how customers interact with the proposition without fundamentally changing Revolut's role as the interface.
Scalable Capital's Agentic Investing goes further. Clients can connect external AI assistants such as ChatGPT, Claude or Grok to the brokerage platform and use natural-language interaction for trading, savings plans, watchlists, price alerts, market data and investment search. Trades and savings plans remain subject to explicit client approval before execution.
Interactive Brokers takes a related but more open approach. Its MCP-based AI Integration allows clients to connect their brokerage accounts to external AI tools for portfolio analysis, investment research, risk monitoring and the generation of trade instructions. Clients retain control over which instructions become orders and when those orders are submitted.
The important distinction is not simply how much AI each broker uses. It is where the AI sits.
Across the wider market, however, the commercial evidence behind these models remains far thinner than the volume of AI announcements suggests.
Most AI activity in brokerage still improves today's business or evolves the existing proposition. That can be strategically and economically significant without creating a fundamentally different brokerage business.
But the boundary is beginning to move.
Scalable Capital and Interactive Brokers demonstrate that investment interaction no longer has to remain entirely inside the broker's own interface. Regulated brokerage capabilities can begin to sit behind external AI interfaces while execution remains under customer control.
That does not yet prove a new brokerage business model.
What it does is make a different set of questions increasingly real: who owns the customer interaction, where distribution begins, what happens to acquisition and servicing economics, and whether brokerage capabilities increasingly become infrastructure used through interfaces the broker does not control.
We do not yet see evidence for a standalone "AI-native broker" category with proven economics. Nor do we see evidence that external AI interfaces will necessarily displace brokers' own customer channels.
We do see a developing boundary between operational improvement, proposition evolution and potentially new forms of distribution and business — and that boundary is worth understanding before committing to either hype or dismissal.
Customer adoption, incremental assets and order flow, willingness to pay, durable cost advantages, regulatory boundaries and the economics of third-party AI distribution remain open.
The strategic significance may be real before the revenue pool is.
Where does AI materially change brokerage economics rather than simply add another feature?
Which capabilities are becoming table stakes, and which could change the proposition or distribution model?
Can AI make customer segments economically viable that are difficult to serve today?
What happens to the role of the broker if AI assistants become meaningful investment interfaces?
We exchange perspectives with financial institutions, founders, technology companies and researchers working on the future of brokerage.
If this theme is relevant to something you are building, evaluating or trying to understand, we would be interested in comparing notes.