Sub-processors

Cloud hosting and managed database, application hosting for the web interface, language and text-understanding services used to read job descriptions and CVs, transactional email delivery, a people-search database where a customer turns on outside search, and — where a customer enables AI interviews — real-time interview infrastructure, speech recognition (turning the candidate’s spoken answers into text) and speech synthesis (the interviewer’s voice).

The people-search database is used only where a customer turns on outside search. It receives a role’s title, the skills it asks for and, where set, a country, and returns professional profiles. We keep only a name, a location, a profile link, a headline and summary, work history, skills, education and a work email: never a demographic, a personal email or a phone.

Speech recognition receives the candidate’s microphone audio and returns text; speech synthesis receives the interviewer’s words and returns audio. Neither is asked for anything about how a candidate sounds — no emotion, tone or accent. The interview infrastructure is set not to record the audio or the transcript. The recording itself is stored by us and deleted on the schedule the workspace sets.

Sub-processors are engaged for a stated purpose and receive only the data that purpose requires. Text-understanding services receive document text and a role’s requirements; they are not given a candidate’s identity in order to decide who is retrieved, because the retrieval half of the product never sees a name.

Notice of additions and replacements, and the right to object, are set out in the data processing agreement rather than left to whenever this page was last edited.