Article Finder now builds a search plan before it pulls results. It defines the likely topic, expands into supporting phrases, filters out common false matches, and keeps articles only when they are contextually relevant to your search.
Use a topic, a person, a company, an event, or a phrase. The search plan will try to separate the intended meaning from irrelevant mentions.
Single-word queries are treated more cautiously, and the planner looks for supporting context before an article is kept.
Infer what the query most likely refers to before searching the news index.
Generate supporting topics, institutions, phrases, and keywords that belong to the same topic.
Reject arbitrary articles that mention the query but do not match the intended subject.