Zalos Brings Human-Like AI Agents to Finance Systems

Zalos, automating repetitive finance workflows, without having to change your finance systems, led by William Fairbairn and Hung Hoang, has raised a $3.6 million in seed led by 14 Peaks with participation from Cohen Circle, 20VC, and notable angels.

Finance teams rely on complex ERPs, yet still manage manual, repetitive work across disconnected systems. Zalos automates these workflows using computer agents, without requiring system replacement.

Modern finance teams operate across fragmented systems that rarely integrate, forcing manual data stitching to complete core workflows like billing, closing, and reporting. Zalos addresses this with agentic software that works across existing tools, automating processes while understanding deeper business context, without replacing the stack.

Computer Agents are emerging as a defining AI shift for 2026, following generative AI, multimodal systems, and reasoning models. While generalist agents from companies like OpenAI and Anthropic are entering the space, Zalos is purpose-built for finance operations, where accuracy and accountability are critical.

Finance workflows demand near-perfect precision, domain-specific capabilities, and fully auditable actions. Still in its early stages, comparable to the era of GPT-3.5, the Computer Agent market is evolving rapidly. Zalos focuses on purpose-built infrastructure and evaluation systems designed to achieve the reliability required for automating finance operations at scale.

Finance teams continue to rely on disconnected systems, resulting in manual execution of critical workflows, noted Zalos CEO and co-founder William Fairbairn. Zalos is built to operate on top of existing infrastructure, using computer agents to execute end-to-end workflows, offering a faster path to transforming finance operations without replacing the underlying stack.

Zalos transforms recorded finance workflows into computer agents that log in, navigate systems, input data, and validate controls across ERPs, spreadsheets, email, and internal tools. The platform operates within systems like NetSuite, Sage, and SAP S/4HANA without requiring heavy integrations.

Each action is recorded in an auditable log, with enterprise-grade security features including SOC 2 compliance, single sign-on, role-based access, and on-premise deployment. Key use cases include billing automation, month-end reconciliation, and cross-system KPI reporting.

Zalos was founded by CEO William Fairbairn and CTO Hung Hoang after independent experiences led to a shared insight. Fairbairn, during his time at Agicap, observed consistent frustration among CFOs around lengthy, high-risk ERP implementations with limited upside. Hoang, after five years at Apple Pay, focused on computer agents as a way to bypass the API limitations that have hindered finance automation.

The company was built following participation in Y Combinator, with a focus on specialized agents designed to replicate how finance teams operate within existing tools.

A structural challenge defines enterprise finance, as innovation in legacy ERP systems has slowed, resulting in increased manual work instead of true automation, noted Zalos CTO and co-founder Hung Hoang. While AI-native ERPs may appeal to companies yet to adopt a system, replacing deeply embedded infrastructure remains unattractive for most midmarket and enterprise teams due to years of operational dependence and the lasting impact of difficult implementations.

The emergence of reliable computer agents introduces a third path: automation that operates on top of existing systems, executing workflows as a human would. Trained once using screen recordings, these agents can run processes continuously with a level of speed and consistency beyond human capability.

Finance operations remain difficult to transform with AI due to deeply embedded, complex systems. Zalos overcomes this by training agents on existing workflows, enabling automation without replacing infrastructure or losing domain expertise.

Zalos plans to expand beyond midmarket ERPs into enterprise and on-premise systems, building a unified context graph across the finance stack. This approach aims to enable CFOs to deploy multiple agents and significantly enhance the impact of finance operations.