Serval Secures $47M to Power AI Agents for IT Service
▼ Summary
– Serval raised a $47 million Series A round led by Redpoint Ventures with participation from other top venture firms.
– The company’s client list includes major AI players like Perplexity, Mercor, and Together AI, which is considered more impressive than its investors.
– Serval uses agentic AI to automate IT service management with two separate agents: one for coding internal automations and another for responding to user requests.
– The system prevents rogue AI actions by requiring managers to set specific rules for each automation tool and not allowing agents to perform unauthorized tasks.
– This approach provides full visibility and control over AI agents while making automation easier and more secure than manual processes.
Serval, an enterprise AI company focused on IT service automation, has successfully raised $47 million in a Series A funding round to advance its unique approach to agentic AI systems. The investment was spearheaded by Redpoint Ventures, with additional backing from prominent firms including First Round, General Catalyst, and Box Group. Beyond the impressive financial support, Serval distinguishes itself through its roster of high-profile clients, such as Perplexity, Mercor, and Together AI, underscoring the practical value and trust its solutions have already earned in the market.
The company’s core innovation lies in deploying two distinct AI agents to streamline IT service management. One agent is responsible for coding internal automations for routine tasks like software authorization or device provisioning. This agent functions as a “vibe-coding” tool, working largely autonomously while remaining under the supervision of an IT manager. A separate help desk agent handles user requests by activating these pre-built tools according to established rules, ensuring consistent and controlled execution.
According to Serval CEO Jake Stauch, the primary goal is to minimize the effort required to create automations. “We aim to make automating a task permanently simpler than performing it manually just one time,” Stauch explained. By eliminating the perceived marginal cost of building these automations, Serval encourages broader and more consistent adoption across organizations.
This two-agent architecture also enhances security and oversight. When an automation is developed, managers define specific rules governing its use, adding a crucial layer of permission control. This design prevents overreach by help desk agents, a common concern with more generalized AI systems. “The last thing anyone wants is for an AI to comply with a reckless command, like deleting all company data,” Stauch noted. “Our system would simply respond that it lacks a tool for such an action, while still being able to assist with approved tasks like password resets.”
Because the underlying tools are deterministic, they can incorporate highly detailed permission protocols. These might include multi-factor authentication requirements or time-based restrictions. Should rule adjustments become necessary, Serval’s AI agent can quickly modify the codebase to reflect new policies.
This strategy addresses a widespread challenge in agentic AI: maintaining visibility and control. “Full oversight into what an AI agent is doing is non-negotiable,” Stauch emphasized. “Serval enables organizations to build their own tools and precisely customize the permissions and approval workflows behind them, putting humans firmly in the loop.”
(Source: TechCrunch)
