The trend towards AI
Although AI use is not yet widespread within payroll, given the transactional nature of payroll processing, we anticipate this will quickly change. We are starting to see organisations explore the opportunities provided by AI, however, this will need to be balanced with managing the risks, given the personal and sensitive nature of the data that a payroll uses, as evidenced by recent payroll cyber-attacks.
The future of AI use by payroll vendors
The Chartered Institute of Payroll Professionals (CIPP) sees payroll leveraging AI to combat fraud, identifying anomalies and for employee queries. These developments will mean quicker payroll processing, reduced risks associated with data errors and fraud, greater expectations on data quality and payroll performance, and potentially reduce the cost of outsourcing payroll. Many of these benefits can be realised using Large Language Models, like the one underpinning ChatGPT. These can compare and analyse large data sets to find anomalies, meaning that payroll vendors can use them to quickly validate payroll information. Large Language Models also have the potential to understand information that is in differing formats, meaning that they can adapt to payroll output reports that must be in a specific format, or cases where the format of a third-party vendor’s information is different to that of a master database.
In the near-term we expect payroll vendors to focus on using AI in data loading and post-processing, as these are easier to automate. For example, after the payroll has run, statutory calculations such as employer tax submissions can be calculated and validated by AI, as data is more static, and rules are more consistent than before processing. In the next generation, a greater number of payroll vendor processes will be automated. This will place an increasing responsibility on payroll vendors to make sure controls are in place over the security of sensitive data throughout the payroll journey.
In the next generation of payroll functions, AI use in case management will be widespread. AI is already starting to be used for this function, for example, we have seen ServiceNow partners using Large Language Models to create naturalistic service bots. In addition, some HCM platforms already have native intelligent advisors which can provide built-in solutions for case management, such as Oracle’s HR Helpdesk.
The future of AI use by payroll clients
In the future, there will be a continuation of the trend towards using AI as a ‘virtual employee’, where AI is given tasks and its performance is managed as any other employee would be. In the payroll space, AI will be used as a virtual employee to further automate data entry, validation, and reconciliation. This frees management to carry out tasks which cannot be automated and allows them to take on a more strategic role. For example, we have worked with a client who has developed an AI tool for automated payroll data validation with Aphy, a company who uses Robotic Process Automation (RPA) to automate repetitive digital processes using ‘digital workers’. Our client’s payroll data was held in several formats as well as in Workday, and the tool allowed data validation across these different formats. In the next generation of payroll functions, RPA use will be widespread, so the responsibility for data entry and validation will fall to virtual employees, and the responsibility for management and strategy will increase amongst human employees.
How we’re using AI to support clients
We support clients to identify AI automation gaps using our leading practice Powered assets, which have AI opportunities built into payroll process flows, reports, and controls. When we work with clients, we also identify specific requirements which we leverage to conduct market research for solutions using Forestreet. We’ve partnered with Forestreet to identify payroll vendors, tools and products to support our clients with integrating AI into their processes. Forestreet offers a SaaS market intelligence and vendor discovery platform to find tools to meet specific needs within your organisation, as well as guiding you to emerging products that are gaining traction in the market that you may be unaware of. Forestreet’s use of AI, analytics, and machine learning broadens the scope of market research as well as condensing the time taken for research to just hours or minutes. Once we have identified your specific pain points, payroll requirements, and the platforms that they must integrate with, we can use Forestreet to identify payroll vendors or automation tools to fit the specific needs of your payroll process, accelerating the timelines to change providers whilst reducing the risk of doing so.