Some startups pride themselves on having prestigious financial backers — but just as important is having prestigious customers.
It’s one of the main points of pride for Serval, an enterprise AI company that announced a $47 million Series A round on Tuesday. The round was led by Redpoint Ventures, with participation from heavy-hitting venture firms like First Round, General Catalyst, and Box Group. But even more impressive than the funders is the company’s list of clients, which include major AI players like Perplexity, Mercor, and Together AI.
In broad strokes, Serval is using agentic AI models to automate IT service management, but the company has a unique approach that takes advantage of agentic AI’s powers while avoiding many of its pitfalls. One agent is used to code internal automations for everyday tasks, like authorizing software or provisioning a device. The founders see it as a kind of vibe-coding tool, overseen by an IT manager, but doing most of the work on its own. A separate help desk agent responds to user requests by calling those tools on command, following rules established by the tool.
Serval CEO Jake Stauch says the key was to make the process of building a tool as simple as possible.
“We don’t want them to feel the marginal cost of building these automations,” Stauch told TechCrunch. “We want to make it easier to automate something forever than do it manually once.”
Splitting the task into two agents — one to build tools and one to use them — also gives managers a way to keep an eye on permissions. When an automation is created, the manager will set rules for when it can be used, which provides an extra line of defense against overeager help desk agents.
Enterprise clients are keenly aware of the risks of a rogue AI system, which is part of why Serval decided against a single all-purpose Help Desk Agent.
“You don’t want someone to go into Slack and say, ‘Hey, I want to delete all the data at the company,’ and the very helpful AI agent responds, ‘Great, I’ll delete all the data,” Stauch told TechCrunch. “Instead it will say, ‘Hey, I don’t have a tool for deleting all the data the company. But I do have a tool for resetting your password or doing one of these other tasks.’”
Because the tools themselves are deterministic, they can include extremely complex permissions, like only allowing certain actions after a certain multi-factor authentication process or within a certain time frame. And any time those rules need to be changed, there’s an AI agent at the ready to dive into the codebase and change it.
It’s a new approach to the very common problem of how to oversee agentic AI systems. “You want to have full visibility and control into what that AI agent is doing,” says Stauch. “And you do that by using Serval to build those tools and customize the permissions and approvals behind them.”
