– Kriti Bansal, Deputy Director of Finance and Accounting, AlphaPoint
Ask an expert
Q. Can advisors work with artificial intelligence to ensure clients are safe from fraud?
A. Yes, but AI should support advisors and not act as an independent decision maker. It can flag unusual wallet behavior, suspicious contracts, phishing patterns, and risky authorizations before the damage happens. The biggest vulnerability today is giving AI agents direct, unconditional wallet permission, and this can make the agent itself a massive attack vector for social engineering or bad on-chain data.
Q. What does real-time security look like in the age of AI?
A. In the age of AI, real-time security must be predictive and proactive, not reactive. Real-time security means pre-signing warnings, continuous monitoring of the wallet, immediate warnings of abnormal activity and blocking of risky approvals before funds can be moved.
Q. How can a money manager automate a defense layer that acts as a continuous threat monitor?
A. Money managers must move away from legacy, externally owned wallets and move to programmable smart accounts such as ERC-4337 or EIP-7702. This transition makes it possible to write automated, programmatic safeguards directly at the account level. They can use automated monitoring of wallets, approvals, contract risks, transaction patterns and exposure limits, with human escalation for anything unusual.
