08/10/2026
Our AI Consultant Leevi Vahvelainen took the stage at NordiCHI 2026 yesterday to share insights from our work on quality assurance for LLM-based solutions. The talk was based on a peer-reviewed industry experience paper by Leevi and Siili’s Lead Designer Amy Järvisalo.
The talk was based on a peer-reviewed industry experience paper by Leevi and Siili’s Amy Järvisalo. The paper explores a practical challenge with LLM solutions: useful behavior, potential failures and acceptable boundaries can be difficult to define upfront. Through the development of an LLM-enabled website search assistant, the work shows how a codesign approach that involves stakeholders directly in testing can help uncover context-specific issues and refine how the system should behave.
Great to see our Siilis bringing hands-on experience into the research community and contributing to the wider conversation around building reliable AI solutions.
The paper will be openly available through ACM: https://siili.pro/4jIAK0G
| | |