04:05 INTERVIEW
Fidelma McGuirk: When all the AI hype started three years ago, we were perfectly positioned to include AI on top of what we had because we had all the building blocks in place. The single most important building block for AI is that you must have content for AI to read. With data being sent from customer sources combined with the data we already had, we developed automated validation tools and then automated reconciliation and variance tools. We focused on creating our global payroll data
model first. Then we built our integration configuration tools. We never wanted integration to be a blocker to adoption. So, by the time AI came on two or three years ago, we were already eight years in the market, organizing and solving all these problems in layered ways that are compatible with AI. As a result of that, we were able to develop really interesting AI tools like Payslip Alpha. For one of our customers, for example, when we were helping them classify their data, we reduced the work time by 75%. We have a payroll data element classification tool and a pay elements data mapping tool, all within our suite. One nice feature is that customers, for example, can ask questions in Japanese and the answer will be in Japanese. It has been very gratifying that the kind of core boring decisions we made at the start about the data model ended up being foundational for incorporating this next advancement. GPA: What do clients often miss, overlook or fail to consider that can cause more work for them or you as you bring them on board?
Payroll people should feel like they have a full license to change a lot of things. If we forget everything that’s in place at the moment, what should the best scenario look like? Give yourself a chance to really become a big system thinker and see things differently.
32 I 04:05
GLOBAL PAYROLL MAGAZINE ISSUE 23
Made with FlippingBook - Share PDF online