First Place at the Triverra Hackathon!
We’ve got amazing news! 🎉
Last Tuesday, 10 February 2026, we participated in the Triverra Hackathon — and we finished in shared first place, together with the team from NovioQ. A well‑deserved congratulations to them as well; they were a strong and incredibly fun team to share the victory with!
We also want to compliment the organizers and the jury. The day was extremely well‑structured, challenging, and above all, a lot of fun and highly educational to be part of.
What did we build?
Our use case focused on a challenge many organizations struggle with:
How do you preserve critical knowledge when experienced engineers leave?
Documentation is often incomplete, and the real expertise — the nuance, the edge cases, the why behind decisions — tends to live in the minds of people who won’t stay forever.
We’ve shown that with modern AI, this process can already be tackled successfully — and that we can even deliver it within a single day.
How was that possible?
Thanks to the OutSystems AI‑Powered Low‑Code Platform and the Agent Workbench, which allowed us to build, test, and integrate with powerful AI models at incredible speed.
GAP Analysis AI – what’s missing?
Huginn reads uploaded documents, notes, and artifacts and analyzes where knowledge gaps exist.
Not based on gut feeling, but objectively: Which parts are documented? Which sections are incomplete? Which themes are entirely missing?
Interview AI – targeted knowledge extraction from experts
Instead of long, chaotic knowledge‑transfer sessions, Huginn conducts focused conversations with experts using a Socratic interview style.
The AI asks exactly the questions needed to fill the identified gaps — and continues until the information is clear. Result:
- No repetition
- No irrelevant information
- Much higher quality and depth
Analysis & Progress‑Control AI – are we complete?
After each interview, Huginn automatically checks progress:
- Which topics are now fully covered?
- Which questions still need to be asked?
- Is any information missing in critical categories?
- Do we need to bring in additional experts?
The tool ensures we work step by step toward complete coverage of all essential knowledge.
Why is this special?
This hackathon proved that with OutSystems ODC, the Agent Workbench, and smart AI models (such as Claude Sonnet & Haiku), a solution doesn’t have to take months — it can be up and running within a single day.
Other companies immediately recognized the challenge — everyone struggles with this.
And the strong interest in our solution confirmed that we created something genuinely valuable.