Automated Accounting: A Efficiency Booster for Finance Teams with Andreas Martin von Flowers

This interview is based on an episode of the torq.partners Finance Podcast featuring guest Andreas Martin, co-founder and CEO of Flowers. The full episode is available on Spotify and YouTube as a video podcast.

What does Flowers do—and what problem does it fundamentally solve?

Flowers focuses on workflows, makes them transparent, and automates them. The goal is to prevent recurring errors, manage handoffs smoothly, and design processes so that technology truly helps in everyday work.

Why are finance processes, of all things, so crucial?

Most breakdowns occur at departmental boundaries. Without an end-to-end view, there is a lack of oversight and accountability; mistakes cost time and money. Transparency comes before automation—simplify first, then automate.

Where is the greatest potential for automation in finance?

When it comes to automating accounting beyond preliminary account assignments: accruals, splits, provisions, and payment allocation—including foreign currencies and credit cards. The degree of automation is crucial: 30% versus 90% makes a massive difference in costs and liquidity.

Why do automation projects still fail so often?

Choosing the wrong tools and relying on piecemeal solutions fragments the system landscape. Furthermore, a holistic approach is too rarely taken: without contextual data such as contracts, budgets, or approval rules, automation remains superficial and vulnerable.

How do you take a pragmatic approach to new projects?

Using proven templates from hundreds of setups. Historical transaction data delivers immediate quick wins and trains the model. The core logic remains deterministic; AI only supplements it where it can increase speed.

What do you mean by task and process mining—and how does it actually help?

It maps out the flow of a process without any gaps. This reveals bottlenecks: for example, approvals that result in lost cash discounts, unnecessary manual steps in standard invoicing, or redundant approval levels for small amounts.

Realistically, what role does AI play in finance processes today?

AI delivers the final few percent: VAT Act checks, context mapping, email extraction, contract summaries. Posting decisions must be 100% accurate, which is why they are deterministic. AI is integrated into the templates on a plug-and-play basis—it’s not a DIY project.

What do teams need to have in order to benefit from this?

Ideally, not much. For high levels of automation, understanding the context and limitations is key: What data does the AI receive, and what is it allowed to do—and not allowed to do? Control remains essential, because blind trust leads to errors.

How do you get teams on board with change—despite initial skepticism?

Involve them early on and make the impact tangible. When people realize that routine tasks are being eliminated, they quickly shift from skepticism to acceptance. Those who work in the process every day know the crucial details and should have a say in the decision-making.

How is automation changing roles in finance?

Repetitive processing of documents is largely a thing of the past. Design, interpretation, tax classification, and quality assurance are becoming more important. Verification is shifting from manual data entry to intelligent anomaly detection and random sampling.

What quick wins do you recommend to get started?

Start with an end-to-end process, such as invoice receipt, including approvals and posting. Use templates, reduce special cases, and ensure quick wins. The time saved is reinvested in the next process, and step by step, the level of automation increases until it reaches “zero clicks.”