Sarah Perez has been a veteran, seasoned reporter for TechCrunch since August 2011. She has developed a deep bench of expertise in storytelling about technology and innovation. Prior to joining TechCrunch, she worked for more than three years at ReadWriteWeb, where she developed her reporting chops. Her extensive experience in information technology spans many different industries. This experience in banking, retail, and fintech startups provides her a unique perspective on the constantly shifting world of technology.
In a recent briefing, aiCIO caught up with Aaron Levie, co-founder of Box and now its chief executive officer, to get his views on artificial intelligence (AI). He highlighted its transformative effect on enterprise software as a service (SaaS). Specifically, he was doubting the notion that AI agents would take over from existing SaaS vendors. Instead, Levie emphasized that smaller startups, which often lack entrenched business processes, are uniquely positioned to adapt to the emerging agent-first paradigm.
After you go through the hard work of creating a new business process, don’t just write the business rules in the wrong or inefficient business logic. Levie made clear that this is critical because the risk of a change happening drastically increases each day. This viewpoint underscores just how crucial it is to ensure stability and predictability in long-running business models.
Levie pointed out that while larger companies may find it challenging to shift existing processes, new startups can innovate by designing their operations around AI agents. “We are in this window right now that we have not been in for about fifteen years, which is — there’s a complete platform shift happening in tech that’s opening up a spot for a new set of companies to emerge,” he noted. This change represents a huge market opening for the startups willing to step into the breach and adopt the agent-first mantra.
As promising as a seamless integration of generative AI could be, Levie warned against ignoring the real risks that come into play here. He stressed the importance of data security and operational integrity in deploying AI agents. “If you’re doing something mission critical — we’ve already seen great examples of either data being leaked because of agents, or an agent going and, you know, maybe blowing up your database or doing something in production you didn’t expect,” he explained. Levie’s testimony underscored the importance of keeping clear lines between deterministic software parts of an application and non-deterministic AI-driven elements.
Additionally, Levie expects a rapid proliferation of AI agents across SaaS ecosystems. “The thing that I’m very convinced of is we’ll have about 100 times more, maybe 1,000 times more, agents than we have people,” he asserted. He sees this resulting in a new level of user engagement with software systems never before achieved.