Assessment quality degrades quietly as items age and models change. Billed monthly with a three-month minimum: each month I re-run a drift check against your scorer, flag items that have stopped discriminating and replace them, run a calibration session with your raters, and QA any new task type before it goes live. You also get a quarterly report and a standing slot for assessment questions, so decisions stop waiting on a consultant.