By admin 
Serval is dramatically reshaping the landscape of enterprise automation with the general availability of Catalyst, its sophisticated AI agent designed for building and managing complex business automations. Officially launching on Thursday, Catalyst is now enabled by default for all Serval customers, ushering in an era where teams of AI agents proactively identify, design, and implement automations. This pivotal development signifies a leap forward from traditional IT service management, positioning Serval as a leader in autonomous enterprise operations.
At its core, Catalyst functions as an administrative "super agent" residing above Serval’s AI-native service management platform. This powerful entity possesses the capability to meticulously analyze a wealth of data, including historical ticket resolutions, established standard operating procedures (SOPs), and natural-language instructions. By sifting through this information, Catalyst can pinpoint recurring tasks and inefficiencies, subsequently drafting the comprehensive automation solutions required to address them. These solutions encompass intricate workflows, specialized skills for AI agents, intuitive forms for user interaction, robust access policies, end-to-end journeys for employee onboarding and offboarding, and insightful dashboards for performance monitoring.
Beyond reactive automation, Serval is also leveraging Catalyst to deploy background agents. These proactive agents continuously monitor connected systems, diligently searching for emergent issues and proposing potential fixes before an employee even has the opportunity to file a support ticket. This forward-thinking approach is crucial in a market where enterprise service management vendors are rapidly converging on AI-assisted workflow creation.
The competitive landscape is heating up, with major players like ServiceNow, Atlassian, and Freshworks all introducing their own AI-driven automation capabilities. ServiceNow’s Build Agent, for instance, translates natural-language instructions into full-stack applications, flows, scripts, and other platform metadata, while its AI Agent Advisor analyzes instance records to identify automation opportunities. Atlassian’s Rovo streamlines the generation of Jira automation flows from plain-English requirements, and Freshworks offers Freddy AI Agent Studio for crafting service agents that operate across its Freshservice workflows.
Serval’s claim to differentiation, therefore, transcends the simple assertion of "using AI to build workflows." Catalyst is engineered as a singular administrative layer capable of orchestrating the entire automation lifecycle. This includes discovering automation opportunities, assembling diverse types of governed automations, and ultimately creating proactive agents that ceaselessly seek out new work to streamline. "You just started with a single prompt, and now you’ve got enterprise-grade workflows ready to deploy that are going to solve all password resets for the entire company," remarked Jake Stauch, co-founder and CEO of Serval, in a recent interview.
From Ticket History to Seamless Automation
A key innovation of Catalyst lies in its ability to analyze existing help desk data before an organization has explicitly decided what to automate. If it identifies a consistently repetitive category of requests, Catalyst can automatically draft the necessary automation to resolve these issues and present the proposed solution for administrator review. Furthermore, users can upload existing SOPs or spreadsheets and instruct Catalyst to transform these documented processes into executable automated systems, significantly reducing manual effort and the risk of human error.
Serval’s comprehensive documentation details Catalyst’s extensive capabilities. It can construct complex workflows, author specialized help desk skills, design seamless onboarding and offboarding journeys for employees, configure granular access-management policies, build informative dashboards, investigate operational anomalies, and even debug failed workflow runs. This marks a significant evolution from Serval’s earlier workflow builder, with Catalyst now intended to serve as the primary interface for configuring the entire platform. The company’s long-term vision is to ensure that any task an administrator can perform through the graphical user interface (GUI) will also be achievable through Catalyst’s conversational and agentic approach.
The underlying workflows generated by Catalyst are code-backed, offering robust and scalable solutions. In a compelling demonstration, Stauch showcased Catalyst’s ability to process a request for password-reset workflows. The agent intelligently identified relevant connected systems, including Okta, Google Workspace, and Microsoft Entra, and proceeded to generate the necessary underlying TypeScript code to perform these actions. Administrators then have the flexibility to introduce approval steps or precisely define which users are permitted to initiate these workflows, ensuring robust governance and control.
Swappable AI Models for Unparalleled Flexibility
Crucially, Serval is adopting a model-agnostic approach to the underlying AI technologies powering Catalyst. Stauch emphasized that the company does not aim to build its own foundational models. Instead, Serval partners with "frontier labs" to leverage state-of-the-art models, conducting rigorous evaluations to determine which models perform best for specific tasks. This deliberate strategy allows for the seamless integration and swapping of different AI models. "You can swap different models in," Stauch stated, adding that Serval also collaborates with enterprises that possess their own proprietary AI models.
Further insights into Serval’s model strategy were shared in a May 2026 interview with Sequoia Capital. Stauch revealed that Serval utilizes models from both OpenAI and Anthropic. He noted that OpenAI’s GPT models have demonstrated superior performance in end-user interactions and tool calling, while Anthropic’s Sonnet and Opus models have yielded the strongest results for the code-generation aspects of Serval’s automation system – a workload directly relevant to Catalyst’s core functionality. Serval maintains a continuous evaluation process rather than automatically migrating all workloads to the newest model releases, ensuring optimal performance and efficiency.
This flexible architecture means that the specific Large Language Model (LLM) is less central to Serval’s competitive differentiation. The company’s documentation now empowers organization administrators to supply their own OpenAI or Anthropic API keys, including compatible custom endpoints. Stauch further confirmed that the broader architecture is designed to accommodate a diverse range of models. However, it is important to note that this does not necessarily translate to a self-service menu for individual model selection for every Catalyst user. Serval’s core value proposition lies in its proprietary "harness" around these models, which includes enterprise context and memory, seamless integrations, expertly generated code, robust permissions, approval workflows, and the critical controls that govern an agent’s operational capabilities.
This code-generation model forms a cornerstone of Serval’s strategic advantage against competitors like ServiceNow. Stauch argues that legacy ITSM deployments often become burdened with custom tables, intricate business rules, and platform-specific expertise, making even seemingly minor automation changes prohibitively expensive. Serval’s approach, in contrast, empowers administrators and business teams to articulate their desired outcomes, allowing the AI to generate the implementation details.
However, ServiceNow is not standing still. Its current Build Agent mirrors Serval’s capabilities by creating applications and code from natural-language prompts, supporting flow design and testing, and operating within ServiceNow’s established governance framework. ServiceNow’s AI Agent Studio enables customers to create agents and agentic workflows, while AI Agent Advisor is specifically designed to analyze operational records for potential automation candidates.
The evolving competitive dynamic, therefore, shifts from simply "who has generative AI?" to a more nuanced question: "how many distinct tools, configuration concepts, and specialized personnel are required to transform an observed operational problem into a fully functional production automation?" Serval’s argument is that Catalyst significantly compresses these steps into a single conversational interface and a more streamlined platform model. ServiceNow, in contrast, offers a powerful yet broader suite of AI and development surfaces, encompassing Build Agent, AI Agent Studio, AI Agent Advisor, Workflow Studio, and AI Control Tower. While this breadth can be advantageous for deeply entrenched ServiceNow customers, it also highlights the complexity Serval aims to simplify. Notably, ServiceNow itself acknowledges that its Build Agent is targeted at administrators and developers who possess the understanding and ability to support its generated outputs.
Atlassian is pursuing a similar trajectory from a different starting point. Rovo can generate "if this, then that" automation flows from natural-language descriptions, while Jira Service Management increasingly supports agents that can triage, investigate, and execute service tasks. Freshworks’ Freddy AI Agent Studio also emphasizes agents capable of end-to-end request resolution, offering pre-built IT and HR agents along with over 30 workflow templates.
Catalyst’s true differentiator, therefore, is not merely the ability to generate an automation from a textual prompt, a capability now shared by many rivals. Instead, Serval’s ambition is to render the entire automation lifecycle itself agentic.
Building Proactive Agents That Anticipate Issues
This proactive approach is most vividly illustrated by Serval’s background agents. Instead of passively waiting for a help desk request, these agents can operate on a scheduled basis across connected systems. They meticulously correlate signals, identify potential problems, and draft remediation plans. In a compelling customer example provided by Serval, an agent analyzed network incidents across two different offices by correlating switch telemetry, DHCP data, and historical ticket information. After ruling out hardware and wireless interference, the agent traced the root cause to configuration drift and generated a remediation workflow for an administrator to review and approve.
"Most AI agents today wait for an employee to ask a question or submit a ticket," Stauch observed. "We believe the future is AI that acts before an employee ever submits a request." This forward-looking perspective underscores a philosophical divergence in Serval’s market positioning. The startup aims to move beyond a model where service management primarily revolves around creating, routing, and tracking support tickets. Instead, Serval envisions a system that proactively eliminates as many requests as possible by transforming repetitive support work into executable automation.
"A lot of the code written in enterprises has nothing to do with software engineering," Stauch explained. "It’s actually internal automations and other scripts for the company, and so we use that technology to build a better service management platform." Serval’s proposition to enterprises is to largely automate these internal scripts. The governance model is paramount, given Catalyst’s ability to generate code and potentially initiate changes across production systems. Serval asserts that Catalyst inherits the permissions of the user operating it and remains scoped to that user’s team workspace. All generated automations begin as drafts, and organizations can implement restrictions on publishing privileges or mandate formal review and approval processes before an automation becomes active.
Customer Data Ownership and Flexible Deployment Options
These robust controls extend to the enterprise data that Catalyst analyzes. Stauch stated that Serval is designed to function as the customer’s system of record, and he affirmed to VentureBeat that "they own all the data." Serval’s Master Services Agreement provides further clarity: customers retain full rights, title, and interest in both their "Customer Materials" – encompassing records, documents, workflows, prompts, inputs, and configurations – and the outputs that Serval generates from them. Serval receives the necessary rights solely to process this information for the purpose of providing, maintaining, supporting, and securing its services. Furthermore, Serval explicitly states that it does not retain or utilize customer materials, inputs, or outputs to train, fine-tune, or improve its own or third-party AI models.
The company’s Data Processing Addendum designates Serval as the processor of customer personal data, allowing processing for operational purposes, responding to support requests, diagnosing issues, and safeguarding the platform. Authorized subprocessors may also be involved. Serval’s acceptable-use terms mandate the maintenance of a current list of AI subprocessors and model providers for customer transparency.
Data residency can also be a flexible consideration. Stauch indicated that customers can opt for Serval as a cloud SaaS service, deploy it on-premises, or host it within their own Virtual Private Cloud (VPC). Serval’s self-hosting documentation now outlines two comprehensive deployment options: a Serval-managed single-tenant deployment within a customer-owned AWS account, or a self-managed deployment on the customer’s Kubernetes cluster, whether in a public cloud or on-premises environment. In the AWS option, Serval commits to operating the installation without persistent IAM access to the customer’s AWS account. This presents two distinct access boundaries for enterprise buyers to carefully consider. Therefore, Stauch’s assertion that Serval "doesn’t touch" customer data should be understood as a statement of ownership and deployment preference rather than a literal claim that the service never processes it.
Early Successes with Ramp and Other Key Customers
Customer deployments provide tangible evidence that Serval’s accelerated build thesis translates into significant operational improvements, though the metrics are primarily drawn from Serval’s own case studies. Ramp, a corporate expense and financial technology firm, reported in a Serval case study that Catalyst has accelerated workflow building by 50%, extending its use across approximately 10 teams, including IT, finance, facilities, people and talent, legal, and business operations. In one notable instance involving a hardware replacement program, Serval claims Ramp automated the replacement of 600 laptops, saving 150 hours of work and leaving only the approval step to human intervention.
Perhaps a more telling example of Catalyst’s impact is what transpired subsequently. After automating laptop replacements, Catalyst suggested splitting Ramp’s shipping logic into separate office and home workflows to minimize errors. Furthermore, employees outside of the IT department at Ramp are now reportedly leveraging Catalyst for analytics, bulk ticket operations, workflow troubleshooting, and HR process automation.
Other Serval deployments highlight the broader operational ecosystem that Catalyst is designed to configure. Mercor has utilized Serval automations to onboard over 4,000 external experts and has expanded the platform’s reach across seven teams. Together AI reports that Serval automates 95% of its just-in-time infrastructure access requests, with robust approval and auditing controls governing sensitive access. Perplexity states that Serval automatically handles more than half of its incoming IT requests and all employee onboarding processes. While these deployments extend beyond Catalyst itself, they effectively demonstrate the comprehensive, cross-system automation substrate that Catalyst is now tasked with building and maintaining. Serval reports that over 90% of its customers adopted Catalyst as their initial entry point for automation during the beta phase. With its general availability on August 20th, Catalyst will be enabled by default for all Serval organizations.
Pricing and the Competitive Standoff with ServiceNow
Serval’s pricing structure is customized based on deployment size and is not publicly disclosed on its website or in its documentation. The company generally offers a single platform fee and typically conducts a pilot program to estimate expected deployment and usage. Stauch indicated that the software license costs can be comparable to those of ServiceNow, but he strongly advocates that the total cost of ownership (TCO) can be substantially lower due to reduced reliance on implementation and maintenance services. "The total cost of ownership is going to be dramatically less – usually half as much, sometimes 10 to 20% of the total cost of ownership of ServiceNow," Stauch asserted. "But the actual software license fee is not necessarily going to be all that different."
Serval’s Genesis and Evolution
Serval was co-founded in 2024 by Jake Stauch and CTO Alex McLeod, both former product and engineering leaders at Verkada. Their entrepreneurial drive was fueled by recurring complaints from IT customers regarding overburdened help desks and the inherent limitations of established IT service-management software. Serval has strategically positioned itself as an AI-native alternative to platforms like ServiceNow and Jira Service Management, aiming to consolidate help-desk ticketing, access management, asset management, and workflow automation within a unified system. Both Serval and Sequoia Capital articulate the company’s core mission as evolving IT software beyond merely recording and routing requests towards achieving automatic resolution. The platform can function as an organization’s primary IT service-management system or augment existing ones with advanced automation capabilities. Its publicly acknowledged customers include Perplexity, Mercor, Clay, Verkada, and Together AI. Serval claims that its customers can automatically resolve over half of their incoming IT requests, with its Together AI case study highlighting the automation of 95% of that customer’s just-in-time access requests.
Investor interest in Serval surged in late 2025. In October, the company announced a $47 million Series A funding round led by Redpoint Ventures, bringing its total funding at that time to $52 million. In December, Serval secured an additional $75 million in a Series B round led by Sequoia Capital, achieving a valuation of $1 billion and raising its total capital raised to approximately $127 million. Participants in this round also included Redpoint, Meritech Capital, and General Catalyst. Serval reported to Reuters that its revenue had experienced 500% growth since August 2025 and that it was expanding its reach beyond IT into operational work performed by human resources, finance, and legal departments.
The Ultimate Test for Enterprise Customers
For enterprise buyers, Catalyst’s most significant test will be its ability to effectively navigate and streamline the automation lifecycle within large, complex, and highly customized environments. ServiceNow has demonstrably advanced its capabilities in generating applications and discovering automation opportunities through AI. Atlassian and Freshworks are also integrating increasingly sophisticated agentic automation into their service platforms. Consequently, Serval cannot solely rely on natural-language creation as its primary competitive advantage.
Serval’s stronger strategic bet lies in its vision of an AI-native platform that transforms the administrative layer itself into an agentic entity. This involves continuously identifying repetitive work, constructing the necessary resources across the entire service stack, making generated code accessible for review, and proactively proposing the next automation before an administrator even opens a workflow designer. If Catalyst can achieve this ambitious scope, the fundamental unit of competition shifts away from individual tickets or even workflows. The true battleground becomes the system that persistently transforms an enterprise’s operational history into novel and efficient automation.
