While some startups highlight their notable investors, having well-known clients can be just as significant.
This is a major source of pride for Serval, an enterprise AI firm that revealed a $47 million Series A funding round on Tuesday. Redpoint Ventures led the investment, joined by prominent firms such as First Round, General Catalyst, and Box Group. However, what stands out even more than the investors is Serval’s impressive client roster, featuring leading AI companies like Perplexity, Mercor, and Together AI.
Serval leverages agentic AI models to streamline IT service management, but distinguishes itself by harnessing the strengths of agentic AI while sidestepping many of its common issues. One agent is responsible for programming internal automations for routine operations, such as granting software access or setting up devices. The founders describe it as a tool for “vibe-coding,” managed by IT supervisors but largely autonomous. Another agent acts as a help desk, addressing user inquiries by activating these tools as needed, in accordance with predefined rules.
According to Serval CEO Jake Stauch, the main objective was to make tool creation as straightforward as possible.
“We don’t want users to feel like there’s any extra cost to building these automations,” Stauch explained to TechCrunch. “Our goal is to make automating a process permanently easier than handling it manually even once.”
By dividing responsibilities between two agents—one for developing tools and another for utilizing them—managers gain better oversight of permissions. When a new automation is set up, the manager defines the conditions for its use, adding an extra safeguard against overly enthusiastic help desk agents.
Enterprise customers are acutely aware of the dangers posed by unsupervised AI, which influenced Serval’s decision to avoid a single, all-encompassing Help Desk Agent.
“You don’t want someone to type in Slack, ‘I want to erase all company data,’ and have the AI agent reply, ‘Sure, I’ll delete everything,’” Stauch told TechCrunch. “Instead, the agent will respond, ‘I don’t have a tool for deleting all company data, but I can help you reset your password or handle other specific tasks.’”
Because these tools are deterministic, they can enforce highly detailed permissions, such as requiring multi-factor authentication or restricting actions to certain time windows. Whenever those rules need updating, an AI agent is ready to modify the code accordingly.
This represents a novel solution to the widespread challenge of managing agentic AI systems. “You want complete transparency and control over your AI agent’s activities,” Stauch said. “Serval enables you to create these tools and tailor the permissions and approval processes behind them.”


