The promise is simple: software that logs into your enterprise resource planning system and does the tedious finance work for you. The execution is anything but. Zalos AI Ltd, a London-based startup, is building computer agents that automate workflows like reconciliation, billing, and month-end close directly inside existing ERP and accounting software [DRJ, Jan 2025]. It is a bet on integration over replacement, and the early signal is strong. The company secured live deployments with major, albeit unnamed, enterprise customers just five weeks after launching its product [Hg Capital, Feb 2025].
The Wedge of the Familiar
Zalos is not asking finance teams to rip out their NetSuite, Sage, or SAP S/4HANA systems [Sovereign, 2026]. Instead, its agents are designed to operate these tools the way a human would, navigating interfaces to execute specific, repetitive tasks. This approach lowers the barrier to adoption by sidestepping complex API integrations and data migration projects. The initial focus is on high-volume, rules-based processes where manual effort is both costly and prone to error. For a CFO's office, the value proposition is operational efficiency without a disruptive platform overhaul.
The founding team brings a blend of operational and technical grounding to the challenge. CEO William Fairbairn was previously the UK general manager at cash flow management platform Agicap and cut his teeth as a consultant at McKinsey [OpenSphere, 2026]. CTO Hung Hoang, educated at the Institut Polytechnique de Paris, provides the technical architecture [LinkedIn, 2026]. Their venture was incubated in Y Combinator's Winter 2025 batch, where they received mentorship from Monzo founder Tom Blomfield [Sovereign, 2026].
The Seed Round and Its Backers
In January 2025, Zalos closed a $3.6 million seed round to fuel its ambitious build. The financing was led by 14 Peaks, a venture firm known for its fintech and B2B software focus, with participation from Cohen Circle, 20VC, and Y Combinator [DRJ, Jan 2025]. The investor syndicate suggests a belief in the team's wedge-and-expand strategy within the massive enterprise finance software market.
The company's early-stage traction, while promising, comes with the standard caveats of a seed-stage venture. The lack of named customer references and disclosed revenue figures makes it difficult to gauge commercial scale and pricing power. Furthermore, the technical complexity of building reliable agents that can navigate the varied and sometimes brittle interfaces of legacy ERP systems is non-trivial.
- Execution risk. The core technical challenge is ensuring the agents operate with near-perfect accuracy across different ERP versions and custom configurations. A single error in financial reconciliation can have significant downstream consequences.
- Competitive landscape. Zalos is not alone in targeting finance automation. It faces competition from established players like OpenClaw, which offers a broader suite of financial operations software, and from large ERP vendors themselves, who are increasingly embedding AI capabilities into their platforms [N24, 2026].
- Expansion motion. The initial wedge is clear, but the path to becoming a platform is less so. The company must prove it can move beyond discrete task automation to managing more complex, cross-system workflows that deliver compounding value.
The Next Twelve Months
For Zalos, the immediate roadmap will be defined by proving out its initial deployments. Success will be measured by renewal rates, expansion within existing accounts, and the public naming of its first flagship customers. Technically, the focus will be on hardening its agents for scale and expanding its library of supported ERP systems and automated workflows.
The $3.6 million seed, led by 14 Peaks with checks from Cohen Circle and 20VC, provides a runway to tackle these challenges [DRJ, Jan 2025]. The question for Fairbairn and Hoang is whether they can convert their promising five-week launch velocity into sustained enterprise momentum. Can an agent that mimics human clicks inside a legacy system become indispensable to a finance department?
Sources
- [DRJ, Jan 2025] Zalos Raises $3.6M To Build Computer Agents That Operate Finance Systems the Way Humans Do | https://drj.com/industry_news/zalos-raises-3-6m-to-build-computer-agents-that-operate-finance-systems-the-way-humans-do/
- [Hg Capital, Feb 2025] Fevered Determination: Building Zalos from Zero to Enterprise | https://hgcapital.com/insights/orbit-podcast/fevered-determination-building-zalos-from-zero-to-enterprise
- [Sovereign, 2026] Zalos raises $3.6M for autonomous AI agents in finance | https://www.sovereignmagazine.com/article/zalos-agentic-ai-finance-seed-round
- [OpenSphere, 2026] William Fairbairn - Zalos | https://nexus.opensphere.ai/zalos/william-fairbairn
- [LinkedIn, 2026] Hung Hoang Profile | https://www.linkedin.com/in/hunghoang
- [N24, 2026] Zalos Secures $3.6 Million in Funding to Compete with OpenClaw | https://n24.com.tr/en/zalos-funding-openclaw-rivals/
- [Y Combinator, 2025] Zalos: Computer Agents for Finance tasks like reconciliation, in your system! | https://www.ycombinator.com/companies/zalos