What can employers and employees do to completely trust AI in the workplace?
There’s little doubt that AI is already having a rapid impact on the workplace. It’s clear that UK businesses are seeing the benefits but in the HR function, alongside the opportunities, there are important questions around privacy, accuracy and accountability yet to be answered.
Its ultimate success depends on how AI is introduced and used in HR, and whether the right guardrails are in place to maximise its benefits without falling prey to the risks. Has its use been carefully planned to solve a specific challenge? Will it support compliance? And how can employees use AI without putting sensitive information or their own security at risk?
The focus now needs to be on making sure AI in the workplace is fit for purpose. So, what should HR professionals be looking out for?
1. Don't implement AI for the sake of it
Since it boomed, companies are starting to look at the return on investment on AI projects, and the results are underwhelming. A study from Studio Graphene found that just 31% of UK companies using AI have experienced a positive ROI.
The key for implementing AI that actually works is always clearly identifying the challenge or experience you want to improve first, then finding an AI tool or system to fit afterwards.
Tusker, for example, has recently implemented AI to identify where customer journeys are not working, allowing the team to spot issues long before they could become a complaint. Rather than waiting for a driver to get in touch when something has gone wrong, AI highlights the problem, helping teams to proactively solve a problem before it becomes an issue.
Identifying the challenge before implementing AI is particularly important where there is a capability gap between AI tools and HR processes. If existing systems are inconsistent or very manual, introducing AI without a solid foundation of data could worsen existing problems.
It’s best to start out by automating a low-risk, repetitive task to test whether AI delivers value before gathering feedback from employees and scaling.
2. What goes in must come out
AI tools are built on existing information, which means their outputs can only be as good as the data inputted into it. If inaccurate information or errors enter an AI process there is a risk that those mistakes will be carried forward into the automated process, and potentially acted on further down the line.
A good example is in payroll detail input. While it may be possible to have checks that flag obvious patterns or problems, organisations can’t automatically assume that all entries are incorrect.
Ensuring that the data fed into AI systems is accurate and reliable is essential to their success. As ever, HR teams have an important role in making sure the data held within their systems is robust before connecting them to AI tools. The stronger the data inputs, the more reliable the outputs, and the more ROI HR teams see on AI systems.
3. Keep humans in the loop
AI is already well established within HR, with a study by The Access Group finding that 73% of HR decision-makers are already using AI at work, compared with just 44% of employees. But as adoption increases, there has to be a laser focus on how AI is monitored.
When implementing a new AI solution, it’s key to balance what aspects are automated and what still needs to rely on human judgment.
Tusker’s customer support chatbot, for example, takes on routine queries but always routes any sensitive or complex cases to human employees, rather than giving automated responses. AI that deals with people needs to have human oversight, empathy and judgement to avoid impacting customer experience.
AI in HR is best suited to specific, defined tasks that are repetitive and straightforward, and clear boundaries must be established around what the system can and can’t do in situations that need sensitivity. Clear policies are needed, to help ensure that AI knows when to step back and a human needs to take over.
AI can deliver real value for HR, but successful implementation is less about adopting the latest technology and more about starting with a clear purpose, ensuring data is reliable and keeping humans in the loop.
Supplied by REBA Associate Member, Tusker
Tusker is the UK’s leader in salary sacrifice cars. Part of Lloyds Banking Group, it has more than 15 years’ experience in offering an affordable way for employees to drive a new, fully insured, and maintained car. Its scheme, which is available to over 1.8 million UK employees, offers a range of options, from pure electric cars to hybrids and even traditional petrol and diesel vehicles. It provides a tailored scheme for organisations’ individual needs.
