Lior Yogev says he’s barely been home in three months.
The FundGuard CEO and co-founder has spent the better part of the year in client meetings, and the conversation at every stop has been a variation of the same thing. “Everybody is now thinking, how do we remodel how we’ve worked over the last 30 years?”
The forcing function is agentic AI and the shift toward autonomous systems capable of taking action across complex workflows without human instruction at each step. In asset management, the implications are massive. Compliance flags that require manual review could be triaged, contextualized, and escalated automatically. Portfolio data that currently arrives in overnight batch files could flow in real time to the decision-makers who need it.
FundGuard’s platform replaces legacy fund accounting systems with cloud-native infrastructure that handles everything from NAV calculations to portfolio accounting to operational automation, in real time rather than overnight batches. The company has raised more than $150 million, counts Citi and State Street among its investors, and now operates across six cities, including Tel Aviv, New York, and London.
The company’s thesis is that large banks and asset managers were using archaic core systems that were expensive, slow, and error-prone. The way this industry has operated “since the 1970s and 1980s” starts to look unprepared for the variety, complexity, and volume that he describes as spiraling out of control. “They can clearly envision the future that’s going to be totally different than the way that they’ve been working the last two or three decades,” Yogev explained.
The first obstacle is the infrastructure itself. The core systems still running much of the asset management industry were built in the 1970s and 1980s and updated minimally since. It is still a conservative and traditional space: large, singular, slow to change. They also process data in overnight runs rather than continuous streams. In Yogev’s framing, this is not an inconvenience to be worked around. “Legacy systems, because they’re [so] monolithic and batch-based, just don’t carry the weight and can’t really interact with the infrastructure that’s being built in the world.”
This creates a sequencing problem that the industry is only beginning to confront honestly. Agentic AI requires live data, modular architecture, and cloud-native infrastructure. Almost everything being built to power the AI revolution assumes these things exist. At most large financial institutions, they do not. The result is an enormous appetite for what AI promises, but paired with the structural inability to capture it.
The second obstacle is governance. Yogev described financial institutions as simultaneously the most excited and the most restrictive audience he encounters. “Most large financial institutions are going to ask you to turn off or disable the use of models, at least at this stage, and not even use their data in an anonymous way.” The concerns relate to data leakage, penetration points for bad actors, and regulatory exposure.
He draws a parallel to the cloud and to institutions' concerns about putting client data anywhere outside their own servers. It felt reckless, legally exposed, and competitively dangerous— until it didn’t. “Just like with the cloud, it took a few more years to be fully embraced by financial institutions. It’s going to be the same with AI,” he predicted. “Everybody is looking at what we’re delivering and gets very excited. And they go with us because they know that we’re futureproof.”
Yogev does not believe the AI transition will move as slowly as cloud, but warns: “If you don’t do it, you’re essentially going to die.”
Institutions that have already modernized their core infrastructure by moving to cloud-native platforms, and in other ways, will be positioned to deploy agentic capabilities more quickly. But those still running batch-based legacy systems will eventually face a two-front problem: they will need to rebuild the infrastructure and close the AI gap, while competing against those who solved the first problem years ago.
The $200 trillion asset management industry has spent 30 years optimizing around the constraints of its technology. It is now up to institutions whether they can reverse that relationship before someone else does it for them.
