London-based Zalos has raised €3.1 million to push a focused idea forward. AI agents that actually do the work inside finance teams. Not dashboards, not copilots, not layers of “insight.” The focus is execution.
Their approach is straightforward on paper. Instead of replacing existing systems, Zalos builds agents that log into them and run workflows end to end. That detail is easy to overlook, but it is where most of the value sits.
Why finance teams are still stuck in manual work
Most enterprise finance teams already have a full stack of tools. ERPs, accounting platforms, reporting layers. On paper, everything should work together.
In reality, a large part of the work still happens manually. Teams download files, move data between systems, reconcile numbers, and fill in the gaps where systems do not connect properly. The issue is not the lack of tools. It is the lack of continuity between them.
Zalos is targeting this “in-between” layer. Their agents behave like human users. They log into systems, extract data, process it, and complete tasks such as reconciliations or submissions. The key constraint is that nothing needs to be replaced. For most finance teams, that alone makes the approach viable.
The real bet: non-invasive automation
A broader pattern is starting to form across fintech and enterprise software. Instead of asking companies to rebuild their stack, new players are working on top of what already exists.
Zalos fits directly into this pattern. Their agents operate through user interfaces rather than deep integrations. This allows them to function across fragmented environments without waiting for clean APIs or perfect system design.
It is a practical decision. Finance stacks are rarely clean. They are full of legacy tools, custom workflows, and edge cases that do not scale well with traditional integration models. An agent that behaves like a user can move across these systems with less friction.
There are trade-offs. Finance requires reliability, traceability, and control. Zalos is addressing this by adding audit trails and oversight into the workflow. It is still early, but the direction reflects a clear understanding of how finance teams operate.
A crowded but focused category
Zalos is part of a broader wave of companies working on finance workflow automation. The space is getting more attention, and the use case is relatively well defined.
Finance operations offer a good testing ground for this type of automation. The tasks are repetitive, the rules are structured, and the cost of errors is high. That combination creates pressure to improve efficiency while maintaining strict control.
This is where AI agents start to make sense as operators, not just assistants.
What this means in practice
If this model works, the day-to-day work inside finance teams will shift. Less time will be spent on manual reconciliation and moving data between systems. More time will go into reviewing outputs and focusing on higher-level decisions.
At the same time, adoption will not move quickly. Finance teams are cautious by design. Trust, compliance, and auditability will shape how and when these tools are used.
Key takeaways for fintech startups
A few patterns stand out from this move.
- Solving around existing systems is often more realistic than replacing them
- The real opportunity is in removing invisible manual work, not adding new features
- AI agents are moving from assistants to operators
- Finance operations is becoming a primary testing ground for agent-based automation
- Adoption depends as much on trust and auditability as on performance
If you are building in this space, positioning matters as much as product. If you want help shaping that story or sharpening your go-to-market, reach out to us at Your Fintech Story.
