Scalable Capital Opens Its Brokerage Accounts to AI Agents like Claude, Grok and ChatGPT
Trending Topics Jakob Steinschaden ● Covered by 2 sources
Scalable Capital now lets ChatGPT, Claude and Grok place trades in customer accounts. It’s the first bank in Europe to open brokerage to mainstream AI assistants this way.
Based on reporting by Trending Topics, Jakob Steinschaden — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Scalable Capital has turned its brokerage into something AI assistants can actually use. Under a feature called Agentic Investing, customers can connect their portfolio to OpenAI’s ChatGPT, Anthropic’s Claude or Grok and issue securities orders by prompt.
At launch, the list of things these assistants can do is broad. They can execute trades, set up savings plans, manage watchlists, create price alerts, search for stocks, ETFs and derivatives, and pull in free news, real-time quotes and historical price data. Scalable’s own analytics tools are in the mix too, so an assistant can run diversification checks, scenario analyses, allocation breakdowns by sector or region, and risk assessments.
The company is clearly trying to make this feel less like a gimmick and more like a workflow. It gives examples that range from a daily newsletter built from portfolio and watchlist holdings to a comparison against a justETF model portfolio, a CSV export of crypto ETP transactions for tax filing, or a savings plan order calculated from a target monthly contribution. An assistant can even help change the limit on an open order or build a portfolio dashboard with positions, performance and a price chart.
Access comes through either an MCP server or a local CLI app. The MCP route is the full-fat option, while the CLI is aimed mainly at people running AI models on their own device. Setup is switched on in the web interface under Profile > Security, and Scalable says it has published step-by-step guides for common assistants at scalable.capital/agentic-investing. There is a catch, though: individual AI assistants can still block certain actions under their own policies.
Control, at least on paper, stays with the user. The feature can be turned off at any time, existing permissions still apply, login and two-factor authentication are required, and trades plus savings plans need approval before they go through. Cost disclosures and key information documents are shown before each securities transaction, notifications continue as usual, and the company says third-party AI apps operate independently and outside its control.
My take — AI-written commentary, not fact-checked reporting
This is either the start of a very convenient mess or the rare case where finance is being honest about who’s really in charge: the human still has to click yes. That restraint matters, because the industry has spent years pretending automation and responsibility are the same thing. They aren’t, and at least Scalable Capital is not dressing it up as magic.
Read more about this at: Trending Topics