Deadlines and reporting
A calendar tailored to the customer profile, plus preparing and submitting reports directly in the tax authority.
AI Accounting: A Reporting Assistant Without an Accountant
AI Accounting is built around an AI assistant and a different interaction model: the user does not need to figure accounting out alone. The product surfaces priorities, explains what matters, and gives a straightforward path to the outcome.
Overview
Instead of "figuring accounting out on your own", the user gets proactive recommendations, clear priorities, and a product that makes the next step obvious.
The core promise is simple: keep accounting logic inside the system, and expose only the parts a non-accountant needs to understand and confirm.
Key complexities
- Domain complexity. Many exceptions and "it depends" rules are hard to explain in plain language.
- Automation vs control. Routine work should disappear, while critical decisions stay reviewable.
- AI transparency. Users need to understand what the system changed and why.
- Workflow cohesion. The experience should feel like one process, not a stack of separate admin screens.
Make accounting manageable for people without professional accounting education.
The product should guide, not overwhelm.
Video walkthroughThe AI assistant in action - auto-categorisation, deadline tracking, and missing document alerts, showing how the product guides a non-accountant through monthly reporting.
What the product does
Taxes due
The tax amount is calculated from reports and data already in the system, so users see what is due without reconstructing the logic themselves.
Data preparation with AI
Transactions are categorised, VAT is checked, and the assistant manages the dataset toward a ready state. Missing documents are flagged before they become a problem.
My approach
Integration over isolation
My primary focus was ensuring AI Accounting felt like a natural part of the user's daily financial workflow - not a separate tool they had to switch into. I worked across business scenarios to identify where accounting context should surface proactively, and where it should stay out of the way.
Clarity as a design constraint
I reviewed every screen and flow with one question in mind: would a business owner without accounting knowledge understand what just happened and what to do next? This shaped how we exposed AI decisions, surfaced missing data, and communicated system actions back to the user.
End-to-end ownership
I led design across the full delivery cycle - from early concepts through developer handoff and market launch - maintaining coherence across the product and keeping the team aligned on the core experience principles.
What this gave the user
Less interpretation
The system turns accounting rules into clear priorities and next actions.
Better visibility
The user stays focused on what is blocked, what is ready, and what needs confirmation.
More confidence
Statuses and explanations help users verify progress instead of guessing what already happened.
One coherent process
Reporting, taxes, and preparation feel connected instead of fragmented across separate tools.
Outcome
AI Accounting launched as a net-new product - accounting and reporting had not existed in Finom before. The assistant made complex financial workflows accessible to SMB owners without accounting expertise, and became a standalone paid product within the platform.
-
First-ever accounting product in Finom
Designed and launched end-to-end, from zero to a paid subscriber base in the first months after release.
-
Report completion 64%
of users who started tax preparation completed it without contacting support
-
Auto-categorisation trust 78%
of AI-suggested transaction categorisations were accepted without manual correction, indicating the assistant's logic was understandable and accurate
