LLM-Based Discovery of Latent Requirements from Stakeholder Conversations: Preliminary Results from Industry
Mithila Sivakumar 1 2 , Martin Lochner 2 , Shiva Nejati 1 , Mehrdad Sabetzadeh 1 · 2026
Read LLM-Based Discovery of Latent Requirements from Stakeholder Conversations: Preliminary Results from Industry by Mithila Sivakumar 1 2 , Martin Lochner 2 , Shiva Nejati 1 , Mehrdad Sabetzadeh 1 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-25867.txt.
Description
Stakeholders rarely state what software they actually need. They describe their work, and the needs hide inside the description. Mithila Sivakumar, Martin Lochner, Shiva Nejati, and Mehrdad Sabetzadeh built LENS in 2026 to surface those latent requirements from interview transcripts. Alongside extracting requirements that are stated outright, the system reasons over what people said together with facts about their organisation's tools and infrastructure. Inferred opportunities become user stories, each linked back to transcript excerpts so a human can audit the leap.
The preliminary industry test used twelve interviews from cybersecurity operations. Explicit requirements came out with an average F1 score near eighty four percent. The inferred ones faced the sharper test: domain experts judged them. Three quarters of the latent requirements were rated as offering useful automation or time saving potential. The results suggest interviews hold a second, unspoken requirements list.
Catalog record
| CREATOR | Mithila Sivakumar 1 2 , Martin Lochner 2 , Shiva Nejati 1 , Mehrdad Sabetzadeh 1 |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.25867