Serval’s Catalyst Pitches AI-Built Workflows as a ServiceNow Replacement

Serval is launching Catalyst with a promise that reaches beyond another assistant for service teams. The startup says the agent can inspect recurring problems in ticket histories, check connected systems’ available APIs, and generate the code, workflows and access controls to automate them. Serval is using that capability to argue it can replace ServiceNow in large enterprises, rather than merely sell alongside it. The immediate constraint is built into the product: an administrator reviews and approves what Catalyst proposes before it goes live.

A proposed workflow, not an automatic deployment

According to CEO Jake Stauch, Catalyst starts with the problem record rather than simply replaying the steps an IT team used to close an earlier request. It identifies repeated issues in ticket history, evaluates what connected APIs permit, then produces automation components including code, workflows, skills, access configurations and permission checks.

The approval step is central to the pitch. Administrators can inspect each step, limit an automation to designated users, and require sign-off from specific people, groups or workflows. Serval separates the agent that creates tools and automations from the agent that uses those tools to resolve employee requests, creating a boundary between making a workflow and letting it run in production.

$127 millionServal funding

Serval raised a $47 million Series A in October 2025 and a $75 million Series B led by Sequoia. The rounds brought total funding to $127 million and its valuation to $1 billion.

$3.88 billionServiceNow Q2 subscription revenue

ServiceNow reported $3.88 billion in second-quarter 2026 subscription revenue and more than $1 billion in annual AI contract value.

Code generation is only part of the replacement test

Serval, founded in April 2024 by former Verkada employees Jake Stauch and Alex McLeod, is targeting ServiceNow’s Fortune 500 customer base. Stauch says the company intends to fully replace ServiceNow rather than add another AI product to the account. Its platform primarily uses OpenAI and Anthropic models through zero-retention endpoints.

Mike Leone, a vice-president and principal analyst at Moor Insights & Strategy, said TypeScript-based automation can be readable, versioned, auditable and built faster than a traditional services engagement. But he cautioned that code generation is becoming standard in enterprise tooling, and ServiceNow already has a code agent alongside two decades of accumulated process, approval and system context.

ServiceNow is pursuing its own agentic service layer

ServiceNow paid more than $2.85 billion for Moveworks and says its Level 1 Service Desk AI Specialist is live with more than 40 customers. ServiceNow also says that specialist handles 80% to 85% of service requests without human interaction, a company-supplied performance figure.

The operational questions Catalyst does not settle

• Generated workflows need an owner when an enterprise has thousands of them and a regulator asks who approved the code.
• A challenger needs an answer for configuration management and change management, according to Leone.
• Serval has not publicly disclosed its exact annual recurring revenue, profitability or detailed customer count.

Serval says its revenue grew roughly 500% between its Series A and Series B periods and that customers automate more than half of tickets, though those are company-shared figures. Catalyst makes its product thesis concrete; the unresolved test is whether large organizations will trust a young vendor with the permissions, approvals and maintenance burden behind operational automations.