{"paper":{"title":"From Chat to Interview: Agentic Requirements Elicitation with an Experience Ontology","license":"http://creativecommons.org/licenses/by/4.0/","headline":"OntoAgent builds an experience ontology to guide LLMs through structured, more complete requirements elicitation interviews.","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Dongming Jin, Linyu Li, Wenchun Jing, Xiaohong Chen, Yaotian Yang, Yuanpeng He, Zheng Fang, Zhi Jin","submitted_at":"2026-05-07T08:03:03Z","abstract_excerpt":"Requirements elicitation interviews are crucial and time-consuming in requirements engineering, but heavily rely on the experience of requirements analysts. Although recent advancements in large language models (LLMs) have created new opportunities to automate this process, existing approaches rely solely on LLMs for free-form chat without taking into account the interview and development experience. That leads to the omission of implicit requirements and redundant questions. Practically, experienced analysts implicitly follow a structured cognitive framework when conducting requirements elici"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"OntoAgent significantly outperforms existing baselines in both elicitation effectiveness and questioning efficiency, achieving a 33% improvement in IRE and a 21% improvement in TKQR.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That experienced analysts implicitly follow a structured cognitive framework that can be automatically extracted into an ontology and that the four operations (ParseUser, ScoreOnto, ReRankOnto, GatePrune) reliably surface relevant concerns without systematic omission of implicit requirements.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"OntoAgent uses an automatically constructed experience ontology and four guided operations to conduct requirements interviews, outperforming baselines with 33% better IRE and 21% better TKQR in website application experiments.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"OntoAgent builds an experience ontology to guide LLMs through structured, more complete requirements elicitation interviews.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"6436e9101ec624b109520dd8b516938fdeae6194fd1406232ddf6f116828b595"},"source":{"id":"2605.05828","kind":"arxiv","version":1},"verdict":{"id":"ea2a3b77-23b7-4c4d-b161-701153a7df46","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-08T09:13:52.757276Z","strongest_claim":"OntoAgent significantly outperforms existing baselines in both elicitation effectiveness and questioning efficiency, achieving a 33% improvement in IRE and a 21% improvement in TKQR.","one_line_summary":"OntoAgent uses an automatically constructed experience ontology and four guided operations to conduct requirements interviews, outperforming baselines with 33% better IRE and 21% better TKQR in website application experiments.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That experienced analysts implicitly follow a structured cognitive framework that can be automatically extracted into an ontology and that the four operations (ParseUser, ScoreOnto, ReRankOnto, GatePrune) reliably surface relevant concerns without systematic omission of implicit requirements.","pith_extraction_headline":"OntoAgent builds an experience ontology to guide LLMs through structured, more complete requirements elicitation interviews."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.05828/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-20T13:42:04.608210Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-20T09:34:09.467502Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T19:31:19.631374Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T13:14:21.159816Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"fff8475c15f5e854158350c9af62ab377fc8b39ec9976025b8a2eb2ab422bfa1"},"references":{"count":31,"sample":[{"doi":"","year":null,"title":"Techniques for requirements elicitation,","work_id":"286f4e35-1d86-4699-85b7-279fea8ad366","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1993,"title":"Proceedings of the IEEE International Symposium on Require- ments Engineering. IEEE, 1993, pp. 152–164","work_id":"4492dd49-0377-4e80-80ad-04873445fc74","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2013,"title":"K. Wiegers and J. Beatty,Software requirements. Pearson Education, 2013","work_id":"856373d2-74e8-43d7-bd3d-1f5e4dd9a5aa","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Requirements elicitation follow- up question generation,","work_id":"82964f6e-4afc-4e51-9a2f-66e81a3ce1d8","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2013,"title":"The state of the art in automated requirements elicitation,","work_id":"ca43517e-e2c9-40b6-82d0-8e604ed833a9","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":31,"snapshot_sha256":"7c64def647fd47d3978d30efd7b995f5427d007f5e5ab3ccd13842e5ec364d27","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}