Supplementing candidate answers

A method for providing candidate answers in a question-answering system, the method further comprising:

• identifying a first selection of one or more electronic files by a first user, wherein the first user is of a plurality of users that are associated with a plurality of projects, wherein each project of the plurality of projects is associated with both indexed question-answer sets of a database and a knowledge graph of the database;
• determining that the first selection is associated with a project of the plurality of projects, wherein the project is associated with a first set of indexed question-answer sets of the database;
• receiving a first question from a first user at a first point in time;
• adding the first question and a first set of candidate answers that are in response to the first question to the first set of indexed question-answer sets such that the first question and the first set of candidate answers are indexed in the database and mapped into the knowledge graph;
• identifying a second selection of one or more electronic files by a second user, wherein the second user is of the plurality of users;
• determining that the second selection is associated with the project;
• receiving a second question from the second user at a second point in time;
• identifying a second set of candidate answers for the second question;
• identifying a mining range for the second question, wherein the mining range includes a number of business hours preceding the second point in time, wherein business hours only include hours between 8:00 A.M. and 4:00 PM in the time zone of the second user;
• receiving a user request from the second user, wherein the user request indicates a specified type of data of the knowledge graph with which to supplement the second set of candidate answers;
• receiving a second set of candidate answers for the second question;
• determining a subset of the first set of indexed question-answer sets to which the second set of candidate answers will be compared that have been gathered within the mining range and are of the specified type of data;
• comparing the second set of candidate answers to the first question and comparing the second set of candidate answers to the first set of candidate answers using natural language processing and Is-a relationships of the knowledge graph;
• identifying, based on the comparing, the first question and the first set of candidate answer as having a correlation with an identified candidate answer of the second set of candidate answers, wherein the correlation satisfies a confidence score threshold;
• restructuring the identified candidate answer to comprise the second question rephrased as a statement, wherein the identified candidate answer is embedded within the statement;
• and
• supplementing the restructured identified candidate answer of the second set of candidate answers with content of the first question and first set of candidate answers, wherein the restructured identified candidate answer is supplemented in a grammatically sound manner by using Is-a relationships of the knowledge graph, wherein the content includes a hyperlink to data of the database regarding the second set of candidate answers.