An Israeli AI startup claims it can predict investors’ financial decisions before they make them, with 10 large brokerages already using its technology, according to its chief executive. Emerging from stealth on 2 November 2026, 42AI says its system analyses customer behaviour to forecast future trades and help financial firms manage risk.
The company has secured $62.5 million (£46.3 million) in seed funding over three years, including money from its founders and Infinity Group. Its technology promises to help brokerages anticipate trading activity.
However, it also raises questions about investor privacy and whether financial institutions could use behavioural predictions to gain an advantage over their customers.
Your Broker Could Predict Your Next Trade Before You Make It
42AI claims its technology can predict which financial instruments investors might trade, whether they are likely to buy or sell, and how much they could trade.
Rather than relying solely on previous transactions, the startup attempts to identify psychological tendencies that influence financial decisions. Its system examines how customers respond to different situations, including activities that might initially appear unrelated to investing.
These include account registrations, customer service calls, money transfers and previous trades. According to the company, its technology examines sentiment, voice characteristics and customer interactions to construct behavioural profiles. It then uses those observations to estimate the likelihood of future decisions.
According to TechCrunch’s report, 42AI spent three years developing its approach using psychological research, neuroscience and other disciplines.
Co-founders Amit Gal-Or, the company’s chief executive, and Ehud Ezra, its chief technology officer, believe underlying personal tendencies can help explain behaviour in unfamiliar situations. However, the system cannot read minds or guarantee anyone’s next decision.
Ten Brokerages Are Already Using the Technology, Startup Says
Gal-Or told TechCrunch that 42AI is working with 10 large brokerages using its system to manage liquidity, margin requirements and hedging exposure. The firms were not identified, leaving their participation unconfirmed by independent customer statements.
The startup also claims one brokerage generated approximately $30 million (approximately £22.23 million) in additional value through dynamic hedging, which involves adjusting financial positions to manage market exposure. However, the company did not identify the customer or provide independently verified evidence supporting the reported benefit.
42AI claims 98% accuracy in forecasting trading activity, although the figure has a narrower meaning than it might suggest.
Gal-Or explained that it measures whether trading flows fall within predicted ranges, not whether the system correctly forecasts 98 out of every 100 individual trades.
He also reported an F1 score of 0.8 to 0.85 for predicting trading direction and volume. This metric balances precision and recall, accounting for false positives and missed events. Neither performance figure was independently verified in TechCrunch’s report.
The founders acknowledge limitations. Unexpected circumstances or advice from friends and colleagues could change someone’s investment plans, reducing the system’s ability to anticipate their decisions.
Can AI Predict Your Behaviour Without Knowing Your Name?
42AI says its technology can forecast behaviour without requiring customers’ names or other directly identifying information.
Instead, the system links records of individual activities to construct behavioural profiles.
Gal-Or says the company’s customer datasets cover almost 30 million people. However, that figure does not establish how many individuals are actively being profiled or what information they have received about the processing.
Removing names does not necessarily make behavioural information anonymous, particularly when records remain connected to individuals who could potentially be identified.
A brokerage expecting customers to sell during a market downturn could potentially adjust its financial exposure before those trades occur. Such activity might support legitimate risk management, but it could also create conflicts of interest depending on how the information is used.
There is no established evidence that 42AI’s brokerage customers have misused the technology or exploited investors.
UK Regulators Have Already Warned About AI Risks
The UK’s Financial Conduct Authority published examples of good and poor AI practices in financial services on 19 October 2026.
The regulator identified concerns about firms overstating AI systems’ accuracy, reliability or capabilities. It also emphasised consumer protection, transparency, accountability and effective oversight.
The publication was not directed at 42AI. There is also no confirmation in the cited reporting that any of the startup’s brokerage customers operate in Britain.
The Information Commissioner’s Office explains that using personal data to predict someone’s economic circumstances or behaviour can constitute profiling. Depending on how information is collected, linked and processed, UK data protection obligations may apply.
42AI’s founders have cited regulation, monitoring and an internal ethics committee as safeguards against misuse. However, their publicly reported explanations do not establish which specific protections prevent brokerage customers from using behavioural predictions against investors’ interests.
For now, the company’s reported performance remains largely based on its own figures.
The bigger question is not simply whether AI can forecast financial decisions. It is whether giving brokerages greater insight into customer behaviour will ultimately benefit investors or primarily serve the institutions managing their money.