Evidence-Based Toxicology evaluation report for "Using Machine-Learning to Facilitate Data Extraction for Human Health Chemical Assessments: A Proof-of-Concept Protocol" (Angrish et al). After major changes from the first round of triage, the manuscript became much clearer on what it was aiming to achieve (testing the performance of machine elements in an evidence mapping workflow) and was therefore advanced to peer-review. The peer-reviewers commented favourably on this manuscript, stating that exploring the integration of semi-automated data extraction tools such as Dextr into research workflows is valuable. However, the reviewers recommended that a detailed analysis plan in relation to validating the performance of Dextr should be added for the research plan to be considered a full protocol.