TL;DR Over the past year, at Atolls we’ve moved from talking about AI to actually using it. A company-wide hackathon in early 2026 saw 100+ employees build 30 working AI prototypes in under four weeks. These products solved real problems, from deal-parsing tools to a semi-automated invoice process. Six finalists presented in Mallorca. The biggest lesson? Making AI useful means pushing people to think bigger about how they can use it and encouraging them to step outside of their comfort zone.
First of all, let’s get something out of the way: yes, you’re probably sick of hearing about AI. You don’t want to read more corporate babble about how company X is ‘innovatively implementing AI’ and company Y is going through an ‘AI transformation’.
That, though, doesn’t diminish its importance. In the tech world, more or less every business is in the midst of trying to figure out how best to use AI: where to embrace it fully, where to be cautious, how to train employees…the list goes on.
That was, and still is, the case for us at Atolls. For more than a year now, AI has been a big topic at the company. We know it has huge potential to help us build a better experience for our users. And, put simply, we know that we have to embrace it to remain competitive.
But how? Where? Who? Deciding how best to use AI at a company of 1000+ people is no small feat. And while we’re still working on it, we’ve come a long way over the past year. There has been a concerted effort across the company to figure out how and where to use it, from running AI exploration hours in which Atollians from across the company share specific examples of how they’re using AI to a hackathon that helped us turn our ideas into reality. Plus, it goes without saying that every day our usage is evolving, so what you read here is really just a snapshot of the bigger picture.
Setting the stage
Around a month ago, at the end of March, more than 800 Atollians gathered in Mallorca, Spain. We were there for a company-wide get-together where we learned about the vision of the future of the company, explored Palma with iPads while completing challenges, partied late into the night, and yes, some of us may have even woken up early in the morning to go for a run with colleagues (every company has a few of those people, right?).
It came as no surprise to anyone in Mallorca that AI was a big topic at the event, and our CEO Gerhard brought it starkly into focus on day one. He made it clear that using AI isn’t really a choice for a company like ours. But the message also wasn’t one of doom, but rather of how it’ll help us all do better work and, ultimately, build better products for our users.
The thing is, that’s easy to say; it still doesn’t get to how to do it. So, next up: our VP Operations and VP Engineering. They came on stage to share specifics – how AI is already being used across the company today and what it means for each of us to become AI super adopters. And that was just the start: during further breakout sessions throughout the day, Atollians shared more details on how they’ve been using AI (the sessions weren’t exclusively about AI, but it was a key topic throughout). The big event, though, came at the end of the day. It was time to reveal what 100+ creative Atollians had built in one month as part of an AI-enabled hackathon.
How a hackathon inspired Atollians to turn their ideas into reality
In January 2026, Atolls CFO Kathrin Nusser made an exciting announcement: everyone in the company was invited to take part in a hackathon. The goal? Use AI to build something that solves a problem they see at the company. The prize? A fully-funded learning experience (such as a trip to Silicon Valley or to our European tech hubs).
Employees had until the end of January to form teams of 1-3 people and submit their ideas. In the end, more than 100 Atollians submitted 43 creative ideas. Now, the race was on: it was time to build. Teams had until the end of February – just under 4 weeks – to build a working prototype of their hackathon entry.
It’s worth stopping here for a moment to consider what teams were being asked to do. Alongside their normal duties, they had to build a complete working prototype of an idea they had come up with just a week or two ago. Some teams had software engineers and ‘technical’ people. But many didn’t. A couple of years ago, it would have been hard to imagine someone without at least some technical knowledge being able to build a working prototype of a product that exists only in their head, in just a few weeks, and doing it all alongside their normal work. Yet the hackathon revealed that is now possible, and also showed us that there are amazing builders in every team at Atolls.
How Atolls enabled builders to build
Everyone was encouraged to take part in the hackathon, regardless of their experience building products or using AI. And in the year or so prior to the hackathon, employees from a variety of teams had already shared specific examples of how they implemented AI on so-called ‘AI exploration hours’. These sessions helped bring everyone in the company closer to real AI usage.
When it came to the hackathon itself, there were three stages: ideate, build, present.
Stage 1, the ideation phase (where people had to submit their ideas), was purely about being creative. No AI knowledge was needed; teams simply had to explain what specific problem their idea would solve, and how.
Stage 2, the building phase of the hackathon, is when specific support was provided. Teams could request help and advice from Atollians with experience building with AI and were given access to whatever tools and resources they needed. The hackathon champion. Rob, also ran a session showing how to build using Lovable to inspire teams who had no prior experience building with AI (although this was just one of the options for tools and teams didn’t have to use it).
Stage 3 is when the shortlisted projects had to present what they’d built on stage in front of 800+ of their colleagues (more on that later).
The key point here is that people were given freedom to choose how to build and the resources to do it. While some companies are cautiously dipping their toes in the water when it comes to AI, at Atolls were being encouraged to jump right in and get wet.
So, what was the outcome?
As mentioned, there were 43 ideas submitted, 30 of which were actually built (some teams dropped out during the building phase).
Here is a selection of some of the things that were created (and remember, teams had just a few weeks to turn their ideas into reality):
- A deal parsing tool that can parse a list of deals from a specific retailer and automatically highlight which deals are the best (based on price)
- An ad optimisation tool to reduce the time it takes to set up display ad campaigns on our sites, while improving the targeting and relevance to users
- A sentiment and trend analysis tool that takes data from our sites and apps and surfaces emerging trends and pain points from our data
- An insights tool that helps teams answer data-related questions quickly and highlights anomalies in our data, massively reducing the need to check multiple dashboards and data sources each day manually
- An AI product discovery tool that uses ‘hidden’ data to surface relevant products and deals based on a natural language search
- A personalised pitch deck builder that helps account managers understand their clients’ goals and needs better, and assists them in building custom proposals
- An invoice calculation, distribution and approval tool that took a process involving many manual steps and multiple teams in the finance department to a single, semi-automated process requiring just one Controller in the loop to verify
- An internal employee search tool that helps people find a colleague to help solve their specific problem by building a knowledge graph of each employee (based on existing messages and documentation)
- An AI shopping companion that evaluates different savings opportunities to guide users towards the best option for them
As mentioned, that’s just a few examples of the products that were built. Every project solved a specific problem someone at Atolls or our customers faced, and it was really impressive to see what was built. The judges had a very hard time deciding which projects to shortlist for the final stage.
Culmination of the hackathon in Mallorca
Six of the submissions were chosen as finalists, and they were tasked with presenting on stage in Mallorca in front of over 800 of their colleagues and four judges (members of the Atolls C-suite). Pretty daunting.
But, and this is absolutely true, it was also inspiring. These six finalists had built incredible products in a very short space of time. So many people said they were amazed at what they’d had seen and had a newfound enthusiasm for building something themselves. And that was the whole point: the hackathon moved us as a company from possibility to reality. While it’s true that many people were already integrating AI into their work and finding new ways to use it, the hackathon took things to a new level. It showed those who weren’t yet working much with it what is possible.
What this shows
One thing became clear: making AI useful isn’t really about the tools. It’s more about whether people feel able to use them in the first place. Giving people time to explore, making it easy to get started, and not overcomplicating things makes a bigger difference than rolling out a perfect setup from the top.