Serval Launches AI Agent 'Catalyst' to Proactively Resolve IT Issues

Serval, an enterprise service management platform provider, has launched its AI agent 'Catalyst' in general availability (GA) and enabled it by default for existing customers. Catalyst is positioned as a 'super agent' that automatically discovers recurring tasks occurring within organizations and takes responsibility for constructing the automation workflows themselves.

IT departments at enterprises have long faced the challenge of helpdesk inquiries accumulating in large volumes, with staff members handling them through manual work. Many companies have adopted workflow automation tools to address this inefficiency, but the task of identifying 'what should be automated' has largely remained a human responsibility. Catalyst is characterized by AI consistently handling this entire process from discovery through construction and is designed as a unified interface for administrators.

Catalyst's capabilities are multifaceted. It reads past ticket histories, standard operating procedures (SOPs), and natural language instructions, then identifies recurring business patterns. Subsequently, it automatically drafts elements such as workflows, helpdesk skills, onboarding and offboarding procedures, access management policies, and dashboards, which are then submitted for administrator review. Jake Stauch, Serval's co-founder and CEO, stated that 'by inputting just a single prompt, an enterprise-grade workflow solving password resets across the entire company can be generated in a deployment-ready state.' The actual workflows are backed by code, and the demo showcased the construction process in action.

Furthermore, Serval is establishing a mechanism for agents that run continuously in the background using Catalyst. This agent continuously monitors connected systems and presents remediation recommendations before issues become tangible and are reported as tickets. In other words, AI detects system anomalies and prepares countermeasures before employees realize they have a problem.

Competing services employing similar approaches already exist. ServiceNow's 'Build Agent' can convert natural language instructions into applications or scripts, while the company's 'AI Agent Advisor' analyzes instance records and proposes automation candidates. Atlassian's Rovo can generate Jira automation workflows from plain English requirements, and Freshworks provides 'Freddy AI Agent Studio' for creating service agents for Freshservice workflows. AI-assisted workflow automation is rapidly expanding across the entire enterprise service management market, with companies aligning on similar capabilities.

Within this competitive landscape, Serval's differentiation claim centers on a design that consistently covers the entire span from 'discovering automation opportunities' to 'constructing multiple types of workflows,' and further to 'generating proactive agents that continuously identify new automation targets' within a single management layer. Additionally, the long-term goal of the company is to enable all operations that administrators can perform through the UI to ultimately be executable via Catalyst as well. The company's strategic direction appears to be establishing a system where AI handles not the 'configuration' of automation but the entire cycle 'from discovery through operations.'