---
title: Agentic AI startup Zalos Raises $3.6M to automate finance and doesn't even need to replace your systems
description: London and San Francisco startup Zalos has raised $3.6M in seed funding to build AI agents that automate finance work by logging into existing systems, backed by the CFOs of FedEx and Tide.
author: Darie Nani (Editor-in-Chief)
date: 2026-03-24T12:44:46.546Z
updated: 2026-06-30T10:52:15.126Z
canonical: https://www.sovereignmagazine.com/article/zalos-agentic-ai-finance-seed-round
image: https://cdn.nanimediahouse.com/zalos-hero-1600.webp
categories: Artificial Intelligence, Startups
content_type: Spotlight
region: London, San Francisco
publication: Sovereign Magazine
about:
  - type: Organization
    name: Zalos
    description: Zalos builds autonomous AI agents for finance operations that log into existing accounting and enterprise planning systems to automate repetitive workflows. Rather than replacing platforms like SAP, NetSuite, or Sage, the agents operate them the way a human would, learning tasks from screen recordings. Founded by William Fairbairn and Hung Hoang, the company is based in London and San Francisco and is part of Y Combinator batch F25.
    url: https://www.zalos.ai
    industry: Artificial Intelligence
    sameAs:
      - https://www.linkedin.com/company/zalos-ai
---

Most finance teams in 2026 are running on a patchwork. There is one system for accounting, another for customer records, a banking platform somewhere else, and spreadsheets filling in the gaps between all of them. None of these platforms were designed to work together, and the people who use them have quietly become the glue holding everything in place.

That means copying invoice numbers from one screen to another. Checking payment records against bank statements by hand. Updating spreadsheets that feed into reports that feed into decisions. It is not glamorous work, but it is constant, and it is expensive.

London and San Francisco-based startup Zalos thinks there is a better way to handle it, and has just raised $3.6 million in seed funding to prove it. The round was led by 14 Peaks Capital, with participation from Cohen Circle, 20VC, and a roster of angel investors that includes the CFOs of FedEx, Tide, and Ada, alongside the founders of Indeed and 11x.

## The problem with fixing finance

The obvious question is why this has not already been solved. The answer sits at the intersection of risk, complexity, and inertia.

The big enterprise planning systems that companies run their finances on, platforms like SAP, NetSuite, and Oracle, are [deeply embedded](https://www.sovereignmagazine.com/article/finance-chiefs-spent-millions-on-erps-they-hate-maximor-s-ai-is-now-fixing-that-without-start-2). Implementations take 12 to 18 months, cost millions, and carry genuine career risk for the CFOs who champion them. When they go wrong, heads roll. So most finance leaders are understandably reluctant to rip out what they have, even when it is not working well.

[Traditional automation tools](https://www.sovereignmagazine.com/article/refold-ai-bets-agents-can-end-the-api-integration-tax) have tried to bridge the gap, but they are brittle by design. They work by scripting exact sequences of clicks and keystrokes, which means a single interface change can break an entire workflow. Maintaining them requires dedicated developers, and the return on investment often does not justify the overhead.

Meanwhile, the software vendors behind these accounting and planning systems can only control what happens inside their own walls. They cannot force a company's bank, its customer management platform, and its internal tools to cooperate. The result is that finance teams remain the human connection layer, manually stitching data across systems that refuse to talk to each other.

As of 2024, 60% of invoices were still being [manually entered into accounting systems](https://www.sovereignmagazine.com/article/nomerra-raises-2m-ai-fund-administration). That number is falling, but not nearly fast enough.

## Recording your way to automation

Zalos takes a different approach entirely. Rather than replacing existing systems or trying to build integrations between them, the company's autonomous AI agents log into platforms with a username and password, navigate screens, enter data, and check it against controls, exactly the way a human finance team member would.

The process starts with a screen recording. A finance professional records themselves completing a workflow, whether that is a billing cycle across multiple systems, a month-end reconciliation, or cross-system reporting. Zalos converts that recording into an agent that can repeat the process independently.

![Zalos founders William Fairbairn and Hung Hoang. Credit: Gemma Day](https://cdn.nanimediahouse.com/zalos-founders-1600.webp)
*Zalos founders Hung Hoang and William Fairbairn  — Photo: Gemma Day*

"Finance teams have the systems, but they are still doing the work manually because the stack is not connected," said William Fairbairn, CEO and co-founder of Zalos. "We built Zalos on the belief that CFOs should not need to rip out their existing stack to adopt the latest in AI. We want to start by sitting on top of what is already there."

Because the agents interact with software visually rather than through rigid scripts, they can adapt when interfaces change. A button that moves does not break the workflow the way it would with traditional automation, because the agent understands what it is looking at rather than following a fixed set of coordinates.

## Built for the back office, not the browser

Finance is not a domain where 90% accuracy is acceptable. A miskeyed number in a reconciliation or a skipped step in a billing cycle creates problems that compound. OpenAI and Anthropic have both released their own computer agent capabilities, but these are generalist tools built for broad use.

Zalos is purpose-built for finance operations. The platform includes finance-specific capabilities like categorisation, reconciliation workflows, and native handling of two-factor authentication inside enterprise systems. Every action an agent takes is captured in an auditable log, and the company holds SOC 2 Part II certification alongside enterprise single sign-on and role-based access controls.

"The rise of reliable computer agents creates a third path," said Hung Hoang, CTO and co-founder of Zalos, who spent five years building Apple Pay products including its Buy Now Pay Later service. "Automation that sits on top of the existing stack and operates it as a human would. These agents are trained once with screen recordings, then the process is automated forever, never taking a holiday, and at a speed and consistency a person cannot match."

## Why these investors

When the people who actually run finance departments at major companies back a seed-stage startup, it says something about how real the problem is.

Mike Lenz spent decades as CFO of FedEx. Ian Sutherland is the sitting CFO at Tide, the UK digital business bank. Long Dinh holds the role at Ada. These are not venture capitalists making a thesis bet. They are the exact people Zalos is building for, and they are putting their own money in.

Paul Forster, co-founder of Indeed, and Hasan Sukkar, founder of AI workforce startup 11x, round out a group that bridges operational finance experience with AI scaling expertise. The institutional side is equally pointed: 14 Peaks Capital is a fintech-specialist fund, and Cohen Circle brings deep financial services networks.

"Zalos is redefining what software means for CFOs," said Nate Pontician, Vice President at Cohen Circle. "Zalos' computer agents do not just assist; they log in, navigate systems, and complete workflows end-to-end. It is not a copilot. It is a colleague."

Emanuele Larocca, Principal at 14 Peaks, added: "What Zalos has built sidesteps that problem entirely. By operating the systems as a human would, training agents with screen recordings, they deliver the true power of finance transformation without asking CFOs to rip out systems they have spent years configuring."

## What comes next

Zalos already works inside NetSuite, Sage, and SAP S/4HANA, and the company plans to expand into larger enterprise systems and on-premise deployments. The founding team, which joined Y Combinator's F25 batch under the mentorship of Monzo founder Tom Blomfield, deployed with enterprise customers within five weeks of launching.

The $3.6 million will go toward broadening the range of systems Zalos can operate inside, building out what the company calls a "context graph" across the finance stack, and ultimately enabling CFOs to deploy networks of agents across their operations.

The broader [finance automation market](https://www.sovereignmagazine.com/article/4-startups-showing-how-ai-in-finance-automation-pays-off-in-weeks) is valued at roughly $8 billion in 2024 and [projected to reach $18.4 billion by 20](https://www.sovereignmagazine.com/article/spendflo-flo-ai-procurement-software)30. Whether Zalos can capture a meaningful share will depend on whether its screen-recording-to-agent approach holds up at enterprise scale. But the founding team's bet is straightforward: the fastest path to transforming finance operations is not building new systems. It is teaching AI to use the ones that already exist.

## Further Context

**Q: How do you automate an ERP system?**
Traditional approaches involve building API integrations between enterprise planning systems or using robotic process automation to script repetitive tasks. Both methods are expensive to build and fragile to maintain. A newer approach, pioneered by companies like Zalos, uses autonomous AI agents that log into these systems with a username and password and interact with them visually, the same way a human employee would. The agents learn workflows from screen recordings and can adapt when interfaces change, avoiding the brittleness that has limited earlier automation efforts.

**Q: Can AI replace finance professionals?**
Current AI tools are not replacing finance professionals but are increasingly handling the repetitive, manual tasks that consume a large portion of their working hours. Tasks like copying data between systems, reconciling accounts, and compiling reports across multiple platforms can be automated by AI agents, freeing finance teams to focus on analysis, strategy, and decision-making. The shift is closer to gaining a digital colleague that handles the administrative workload than a wholesale replacement of human expertise.

**Q: How do you fully automate finance workflows?**
Full automation of finance workflows has historically been difficult because most companies run their finances across multiple disconnected systems. Accounting software, banking platforms, customer databases, and spreadsheets each hold pieces of the picture, and none were designed to work together. Emerging solutions use AI agents that sit on top of the entire stack, logging into each system independently and completing end-to-end processes like billing cycles, month-end reconciliations, and cross-system reporting without requiring any of the underlying platforms to be replaced or integrated.

**About Zalos**

Zalos builds autonomous AI agents for finance operations that log into existing accounting and enterprise planning systems to automate repetitive workflows. Rather than replacing platforms like SAP, NetSuite, or Sage, the agents operate them the way a human would, learning tasks from screen recordings. Founded by William Fairbairn and Hung Hoang, the company is based in London and San Francisco and is part of Y Combinator batch F25.

[Website](https://www.zalos.ai)