One day after Berlin, the laptops were open again. This time at Personio HQ in Munich.
On 8 October, Light and Lovable joined Personio for stop number fourteen on our global Finance Hackathon tour. A huge thank-you to the Personio team for welcoming us into their Munich HQ and making it a home for a day of finance building.
The brief stayed simple: bring a problem from your working day and build an app to help solve it.
Around the tables, finance people worked through the questions they knew well. How does a hiring decision change our runway? Which overdue customer should we chase first? What actually explains the gap between this month's result and the plan?
By the time the demos started, those questions had become tools people could click through, challenge and keep improving.
The judging panel
The shortlisted teams presented to Tu Nguyen, CFO at Personio, James Arch, Director of AI at Personio, and Sam Landsberg, Head of Sales at Light. Finance, AI and the problems customers bring to us every day, all represented on one panel.
Thank you to Tu, James and Sam for sharing their time and perspective with the builders. And special thanks to our host partners at Personio: opening up your HQ and bringing your finance and AI expertise into the room helped make this day possible.
The shortlisted builds
Four prototypes made the shortlist, using a mix of Light data and demo inputs to explore very different parts of a finance team's day.
Predictive Runway Guard
A runway number is only as useful as the assumptions behind it. Predictive Runway Guard put those assumptions on screen: cash, monthly costs, revenue growth, collection delays and planned hires.
Change a hire's start date, add a cost spike or model a funding round, and the projected cash curve moves with it. Light invoice data supplied revenue and receivables, while editable inputs let the builder test different starting points. The hiring roadmap used example roles for the prototype.
The result brought runway and break-even into the same conversation: how long the cash lasts, when revenue catches up with costs, and which decisions change the answer.
Predictive Runway Guard
Invoice Builder
Invoice Builder focused on what happens after an invoice goes out. Its receivables dashboard brought together outstanding balances, overdue invoices, collections and days sales outstanding, with views across legal entities and currencies.
The brief went further than knowing who owed what. It called for a collections tracker that could show the customer's reminder history, replies and next follow-up, alongside payment terms and payment behaviour. The aim was to give finance enough context to decide what to do next without piecing it together from an inbox and a spreadsheet.
Invoice Builder
Capital Command Center
Capital Command Center organised working capital around three management questions: where do we stand, do we have a problem, and what should we do about it?
Its receivables view compared days sales outstanding with a target, highlighted overdue balances and ranked the customers to chase first. A scenario showed the potential cash released and change in DSO if those customers paid.
The reminder workflow kept the decision with finance: review the history and proposed message, then choose whether to contact the customer, the sales owner or both. The prototype recorded reminders for review; email delivery was a next step. It made the route from a dashboard finding to a follow-up much clearer.
Capital Command Center
Controlling Workbench
Controlling Workbench tackled the work behind the monthly management report. Built by Christian Waldow, it started with a P&L and a visual comparison against plan or the previous year, with filters for company, cost centre, account and month.
From there, a controller could open a deep dive into an unusual variance. An AI agent proposed an explanation, and the person reviewing it could question, correct or confirm the commentary. Confirmed comments fed into the management report, while agent notes kept context for future reviews.
The prototype used company and account data from Light alongside demo postings and plan figures. It also supported German and English, making the same analysis available in either language.
Controlling Workbench
And the winner is… Controlling Workbench
Controlling Workbench took the win in Munich. It connected a familiar sequence of jobs: spot the variance, investigate it, agree the explanation and bring it into the report.
The person responsible for the numbers stayed involved throughout. The build gave them a place to work through the explanation and carry it forward, instead of starting that conversation again each month.
A global community of finance builders
We run these hackathons because we are building a global community of new wave finance leaders. People who don't just report, but build.
Munich added four more examples of what happens when people who understand the work get the tools and space to improve it. Runway, collections, working capital and controlling: everyday finance problems, with prototypes to take back and develop further.
Want to join the next one? Explore the Finance Hackathon tour and follow Light on LinkedIn for future dates.
