

Scalable Capital has opened its brokerage platform to ChatGPT, Claude and Grok, letting clients place trades, build savings plans and check their portfolios by typing prompts into an AI assistant rather than clicking through an app. The Munich-based bank, which holds a German banking licence, says it is the first bank in Europe to do this across all three major assistants at once.
The feature, called Agentic Investing, went live on 25 August. It sits behind a toggle in the app's security settings. Clients link their Scalable account to an assistant of their choice using the Model Context Protocol, a connection standard that Anthropic originally built and that OpenAI, xAI and others have since adopted. Scalable also offers a command-line tool for people running AI models locally rather than through a cloud assistant.
What the assistant can actually do
Once connected, an assistant can search for stocks, ETFs and derivatives, pull real-time prices and historical data, and draw on Scalable's existing portfolio-analysis tools for things like diversification checks, sector exposure and risk scoring. Clients can ask it to compare their holdings against a reference portfolio, export a year of crypto ETP transactions for tax purposes, or work out the monthly contribution needed to reach a savings goal and then draft the order for it.
The company is careful to frame this as assisted investing rather than autonomous trading. Every trade or savings-plan order still requires the client's approval before it executes, deposits and withdrawals can only be made through the app or website, and the AI cannot move money on its own. Clients also have to log in with their usual credentials and confirm two-factor authentication at set intervals, and they can switch the feature off whenever they want. Standard pre-trade cost disclosures and the usual trade confirmations still apply, just as they would for an order placed by hand.
Scalable's chief product officer, Alexander Seipp, told reporters the idea is that a client can start the process of deciding what to invest in inside an AI assistant and finish it inside Scalable's regulated systems. "It certainly creates a level-playing field," he said, pointing to how widely available computing power and AI have become. He expects take-up to vary by client group, though he believes a sizeable share of customers will want to try it.
Why hand trading decisions to a chatbot
The company is explicit that it has no formal partnership with OpenAI, Anthropic or xAI. It is simply building on MCP, an open protocol that any AI provider can plug into, rather than negotiating a bespoke integration with each one. Scalable also stresses that whatever an assistant suggests is not investment advice from the bank itself, and that the client is on their own if they act on it, since the assistants run as independent third-party applications outside the bank's control.
That caveat matters given what independent testing has found about AI trading judgment. A June research project by Elm Wealth had Claude, ChatGPT, Gemini and Grok read historical front pages of the Wall Street Journal and predict the market's next move, without telling the models what actually happened. Claude beat human participants in 76% of roughly 200 sessions and ChatGPT in 63%. Grok and Gemini did notably worse. But the researchers found all four models tended to size their bets far too aggressively given how confident they should have been. They understood risk-management concepts such as the Kelly criterion in theory but applied them poorly in practice. Seipp acknowledged the uncertainty directly, saying it remains to be seen whether clients will find strong returns at low risk by working with an AI assistant.
Part of a wider AI push
Agentic Investing follows Insights, an in-app AI chatbot for financial questions that Scalable introduced in August 2025 and which it also describes as a European first for a bank. The company, founded in 2014 by Erik Podzuweit and Florian Prucker, has more than €60bn in client assets and over a million customers, mostly in Germany and Austria, and raised €155m in June 2025 from investors including Sofina and Noteus Partners.
Podzuweit called the launch "the greatest technological shift in financial technology since internet banking," arguing it sets a benchmark for how people, AI systems and markets will interact. That is a big claim for a feature that, for now, still requires a human to approve every single order. The more interesting question for European fintechs watching this launch is whether retail investors actually want their trading decisions filtered through a general-purpose chatbot, or whether the appeal turns out to be narrower: automating the research and admin around investing while leaving the actual calls to people.
What did you think of this story?
Written by
John Pulisic
92 stories on European Startup Insider
Have a story like this? Pitch it to us.
We publish founder profiles, company news and practical guides from across Europe.
More stories like this
Proxima Fusion Raises €411M, Becomes Europe's Best-Funded Fusion Company
Munich-based stellarator developer Proxima Fusion closed a €411M Series A2 round, reaching a €2.4B valuation. Led by XTX Ventures and East X Ventures with backing from RWE and Google, the capital accelerates development of its Alpha demonstrator and first commercial power plant.









