Thank you very much, Alexandra.
The IBM Phoenix pay system issue is one of the toughest ones, and I've served in cabinet in two governments. This is one of the toughest public.... Well, it's not just public policy. It is the execution issue that I have dealt with.
You mentioned the need to hire pay experts. Part of the challenge is that 700 pay advisers were let go prior to the system actually being operable and the new system implemented. One of the keys in this that you learn in terms of the future is that you don't gut the legacy system before the new system is operating. In fact, in terms of some of the people we are hiring, we're actually trying to bring back some of those legacy people, because they have invaluable experience and an understanding of the pay system on a department-by-department basis.
We have been working closely with the union leaders. In fact, we had a discussion with them again yesterday. One of the recommendations they've made, with which we agree, is that we need to put more people back in the departments, so we are working to that end. Yaprak can speak to some of where we are investing, but we're investing $142 million to recruit, hire, and train people across government.
Beyond that, it's important to remember that the IBM Phoenix pay system was conceived about a decade ago. The technology available then and the IT methodologies practised then are very different from those that exist today. While we work very hard to stabilize the existing system with departments across government, working with PSPC and OCHRO at Treasury Board, we also believe that looking at new approaches and taking a look at.... If this were a greenfield project today, if we were looking at this situation as it is today, unimpeded by the blinders that often occur in government around the tyranny of sunk costs—you're trying to make what you've spent money on work—if we were to look at this with fresh eyes, with a fresh team of people, using modern digital protocols and technology, we may find that there is a new way that can actually address this issue faster.
You need to take look at this in a two-track approach: continuing the work we're doing right now to stabilize the existing system, but also at the same time being open to completely new approaches that reflect modern digital today and that weren't even available 10 years ago when this system was conceived.