TECHNOLOGY
an app, and it’s automatically added to the payroll. This has happened automatically because the pay professionals understood what needed to happen and which scheme applied to each employee, allowing everything to operate automatically. In effect, the pay professional has become the programmer.
automated steps the AI software has completed. Knowing what's been done is imperative when AI takes over a human task. Whenever possible, this is presented to both the pay professional and, where applicable, the employee (e.g., salary pro- rata documented on a payslip). 2. It allows the payroll operator to start or stop the AI automation at any point. 3. Payroll operates in a complicated networked environment: information flows bi-directionally between HMRC, pension providers and bookkeeping software. Things can go wrong, and exceptions will occur. For this reason, exception handling always reverts to the pay professional. With complete information, they can review what's happened and determine the best course of action in relation to any exceptions. 4. The final control is to enable the payroll operator to override any aspect of AI automation. This ensures corrections can be made if the AI output is incorrect. How AI might evolve in the future Data correction: One nice feature of AI is that it can correct bad data. An example of this is when starters or employee changes are uploaded from a file. For instance, if a new department name is misspelt, the upload will generally cause an error because the software merely looks for an exact match. However, AI, armed with knowledge of existing department names can find all near matches, just as a human would. If a single close match is found with sufficiently high confidence, the data can be automatically corrected. Another area, which is the most exciting, is using AI for learning payroll. This could be used for new apprentices entering the industry, but it could also be used for pay professionals transitioning from manual payroll to AI-based payroll. n
“In effect, the pay professional has become the programmer”
The new skills pay professionals will need to possess Pay professionals instinctively manage multiple payroll workflows, including payroll approval, along with HMRC and pension submissions. In the AI world, when a pay run is approved, it launches numerous hierarchical workflows: a pay run workflow and sub-workflows to send the FPS, the employer payment summary, pension data file and bookkeeping data. Just like today, these workflows sometimes fail to complete if the internet is unavailable or HMRC or NEST aren't responding. So, being able to understand and handle hierarchical workflow exceptions becomes a new skill. AI also enables new capabilities in payroll. One of these is live data analytics, so even before the pay run has been approved or completed, analytics are available from day one. This is a new feature which can have a transformational effect on how businesses view and analyse payroll. This means pay professionals will also need to become proficient in Microsoft Excel, pivot tables and Power Query. Pay professionals can maintain control over payroll processes and see the changes made by AI AI in Paiyroll provides for four essential controls: 1. It provides complete clarity on the
65
| Professional in Payroll, Pensions and Reward |
Issue 115 | November 2025
Made with FlippingBook - Online magazine maker