FEATURE TOPIC
“Technology can work to get rid of the mundane tasks, freeing resources for professionals to do the complex work”
professionals. In no small part due to the work of the CIPP, payroll departments are getting increasingly louder with senior decision makers in business. That comes from the insights we can provide for strategic decisions. Technology is providing tools to support this, but we still need experienced professionals to interpret and provide insight into the data available. I think we’ll also see a move towards more strategic tasks, as current processes are increasingly automated and AI deals with basic queries. CW: We’ve already seen a change in what we’re looking for when it comes to skill requirements for reward professionals. Gone are the days when we were looking for data input skills, with the introduction of time and absence systems and electronic expenses systems. We now see the skills change to analyst, compliance specialist – the more complex work which requires the human brain to make decisions. In which ways do you see relationships with technology providers evolving for payroll functions? MA: Solutions need to solve problems. In many ways, the software and technology providers direct the conversation towards their products or innovations, but if these innovations don’t solve problems professionals face, they’re unlikely to gain traction. Payroll functions will continue to invest in what works, saves time and results in accurate payroll. With that said, payroll software has been through a rough period of quick government and policy change, and this has directed a lot of effort towards keeping the compliance side of things up to standard. This doesn’t look to be slowing down any time soon. TJ: As a payroll bureau, we’re increasingly invited by software developers and pension providers to participate in Think Tanks, consultations and research studies. This growing collaboration highlights the evolving role of payroll professionals in shaping industry solutions. We’re also actively involved with shaping areas of our current payroll software solution. Traditionally, payroll professionals were software users. Now, we’re script testers and active contributors. From my experience, at present, payroll software providers are focussed on robotic process automation and improving user
experience, streamlining processes to reduce manual input. However, AI is rapidly gaining traction, particularly in enhancing payroll helpdesk systems through generative AI-powered chatbots and online knowledge centres. The demand for greater connectivity between payroll, HR, finance and benefits platforms is gaining traction, reducing instances of data silos and duplication. Seamless data integration – allowing employees to update personal details directly within a centralised system – is already in demand from end users. AM: As competition increases in the market, it feels like technology providers may be driven to free up time for payroll professionals or bureau professionals by having the system take on more of the basic data entry and integration with other partners such as human resources (HR) and HM Revenue and Customs. This then enables clients / teams to raise the profile of payroll, as we get more involved with supporting the business more strategically. JM: There needs to be close engagement with technology providers to drive the solutions we need to our specific problems. I’m sure I’m not alone in seeing that people outside of the payroll industry still tend to have something akin to the stereotypical view that payroll is just pushing a button once a month. Without collaboration from the experts, there’s a real danger that technology providers underestimate the complexities of payroll, leaving us with solutions which don’t add value, at best, or are just plain wrong, at worst. A collaborative approach with suppliers and broader awareness (which the CIPP again promotes), with measures such as responding to government calls for evidence, are needed to ensure the technology meets our requirements. What are the biggest risks and rewards of AI-driven payroll automation? MA: Data security is, or should be, a major concern for companies with regards to their payroll data. AI systems need to be trained. If they’re only trained with internal data, the system is likely to be slow and inaccurate for all but the largest companies with huge
data sets. How do we effectively scale up AI assistance, such as error checks, without training models on external data? The risk of this is that companies may not be willing to let their data train models for other businesses, which is understandable. The benefits could be huge but, in my opinion, they’ll never fully replace the human expert. TJ: The sophistication of cyber threats is a growing risk. AI-generated deep- fake images and phishing emails are becoming more convincing, placing additional pressure on payroll teams to detect fraudulent activities. Strengthening cybersecurity measures will be essential in mitigating these threats. Unfortunately, I can see there’ll be a lag between education and awareness and the next generation of AI threats. Therefore, risk management and mitigation will be key. AM: If we remove too many tasks which need human intervention, the question will arise as to whether payroll department headcounts at the current levels are still required. Another potential risk is removing the ability for new payroll professionals to get involved with basic data entry and query resolution, which is the area I feel you need to start in when first joining the industry to support your learning and career development. JM: Developments in AI open up the prospect of automating tasks which have been too complex to take out of the hands of an experienced payroll professional. However, payroll is an area where nothing less than 100% accuracy is ever truly acceptable, and I think it will take time to really test and ensure any solutions can guarantee that. There’s also the question of the methodology that’s used within AI models for how they make decisions. If It’s something that’s trained on previous decisions rather than clear-cut rules, then there’s the risk of reinforcing the bias of previous decisions. CW: Technology isn’t human; it will only do what we teach it to do. So, the risks are that we may have to reprogram tasks where change happens. Also, if the technology encounters error it could stop, and then a human would need to step in to restart the program. n
29
| Professional in Payroll, Pensions and Reward |
Issue 109 | April 2025
Made with FlippingBook - Online magazine maker