To clarify, there are approximately three million applications made in a year for immigration services, and the oddity is that individuals, for each individual application, repeatedly enter the same information again and again and again, instead of having a central repository for that individual's personal information, which is updated as time goes on. The savings come from the reduced demand for services, the reduction in confusion, the reduction in error rates by individuals, the simplification of the administrative process behind the IRCC or Wizard of Oz curtain. You can have, say, fewer paid employees doing more with less. That's the goal there.
Related to the previous question, I can't resist that we can collect more evidence-based data two ways. One, given that immigration computers do talk to the tax computers, we can track people, say, over the previous five years or over the forward five years and know where they are now. If they enter Canada under the spousal category, we can objectively measure, within a five-year period, how many remain married. That way we know objectively whether there's a signal, good or bad.
The other way is to track the officers. We can track individual decision-maker results the same way. That way, for some officers who may be generating above average refusals—marriage of convenience decisions—we can provide additional training, as suggested by the other witness.