A user inputs a speech including a keyword via a speech input device; a first processor searches a job history by the keyword, the job history being stored on a storage, the job history including a job record, the job record including a set of values having ever been used for a job executed by an image processing apparatus. A job record specifying device includes a second processor that conducts an analysis on different values in multiple job records; selects a speech with reference to the different values; transfers the speech to a speech generator; and finds a specific job record from the multiple job records using a keyword extracted from a speech inputted via the speech input device in response to the speech outputted by the speech generator. The image processing apparatus reflects a target set of values in the specific job record, to the setting of a job.