SharpLink CEO Joseph Chalom sees AI agents moving from financial assistants to autonomous managers of consumers’ money within five years, as users gradually hand over more decisions to softwa
SharpLink CEO Joseph Chalom sees AI agents moving from financial assistants to autonomous managers of consumers’ money within five years, as users gradually hand over more decisions to software.
Speaking with Scott Melker on The Wolf Of All Streets, Chalom said the transition should begin with low-risk tasks before agents gain authority over portfolios.
“Then let them touch your idle cash, put it in higher yielding accounts, then share with it your risk tolerances and your goals and suggest that it build a better portfolio,” Chalom said.
Users could initially review recommendations before allowing agents to execute them, eventually letting them “rebalance your portfolio” and handle insurance renewals and mortgage checks.
Chalom said the relationship would develop as agents learn users’ preferences and users become comfortable with them. More consequential actions, such as lending money or taking on debt, would follow.
That progression could also extend to crypto. AI agents can be given dedicated blockchain wallets and access to smart contracts, allowing them to hold digital assets, execute transactions and interact with DeFi protocols without waiting for a human prompt. Crypto platforms are already beginning to build products around that model.
Coinbase introduced Coinbase for Agents in June, allowing AI agents to trade and make payments through a user's Coinbase account within preset limits. The company says agents can already rebalance crypto portfolios, monitor idle funds and execute trades based on user-defined rules.
Kraken has also been rebuilding its app around embedded AI agents that can monitor markets and execute trades, while Circle's Arc blockchain is designed specifically for financial markets and what the company calls "agentic economic activity."
That gives a practical dimension to the guardrails Chalom described. An agent can be given defined permissions, spending limits or a restricted pool of assets rather than unrestricted access to a user's entire financial life.
“Over time, when your agent learns your preferences, it'll act autonomously on your behalf within guardrails,” Chalom added.
The idea of software acting on a user’s behalf is already becoming a focus for the financial system. Federal Reserve Governor Christopher Waller said Sept. 29 that agentic AI can execute tasks on behalf of people and companies, including in payments. Waller also pointed to questions around authorization, security and trust as agents take on more responsibility.
Those concerns could become more important as agents move from making recommendations to executing financial decisions. IBM says AI agents are already being used in areas including trading, compliance, reporting and risk management.
Chalom expects users to become comfortable with that autonomy relatively quickly.
“And in five years from now, we're not even going to realize what they're doing,” he said. “They're just doing it on your behalf behind the scenes.”
He added that agents should ultimately work for users rather than a financial intermediary seeking to control the relationship.
Chalom joined SharpLink after two decades at BlackRock, where he held senior roles spanning technology, digital assets and strategic partnerships.
Related: Legendary trader who nailed Bitcoin's 2018 crash makes bold new call
