An HR Manager in a large organisation recently recounted how one of their employees had asked ChatGPT to size their role and tell them how much they should be paid. The HR Manager had to gently explain that ChatGPT had misinformed them, suggesting their role was bigger than the Chief Executive’s and that, as a result, the pay level they were seeking was inappropriate.
A Finance Manager in a small not-for-profit organisation asked one of the AI tools how much she should be paid, given her role, responsibilities, and the organisation she worked for. Using publicly available data, the answer was in the ballpark for a Finance Manager in a small organisation. But it didn’t reflect what this Finance Manager was actually being paid. As a not-for-profit, the organisation couldn’t afford to pay prevailing market rates. It attracted people who wanted to give back to society and were happy to work for less than the market rate.
AI may be able to tell you the prevailing rate of pay for a role with the same title, but it can’t tell you what is relevant for your organisation.
AI could even write a position description for a role with the same title. However, it may struggle to make that description relevant to your organisation, or to determine what the role should look like in your specific context.
Decisions about pay levels are often made with very little evidence, or none at all. A hiring manager might decide based on the level of pay a candidate is asking for. Many organisations take a more structured approach to setting pay levels, perhaps by calling colleagues or recruitment firms, using publicly available data, or checking what SEEK indicates a role with that title is paid.
With AI tools that can pull together vast amounts of information and analyse it in seconds, wouldn’t that provide enough information to make a robust pay decision?
Recent Boston Consulting Group studies suggest that AI will reshape more jobs than it replaces, with 55% of jobs expected to be reshaped in the next two to three years. Many employees will remain in roles with the same title but face radically different expectations for how they work and what they produce. Job titles will provide an increasingly unreliable summary of what a job involves.
Deciding pay based on title alone, or even on a brief description of responsibilities and required skills, risks overlooking the real value of the role, let alone the value of that role within your organisation. Analysing the role in terms of its accountabilities and required skills, within your context, helps ensure the organisation reflects those requirements in the pay it offers for the role.
The process of job evaluation considers the requirements of the position and its contribution to the organisation. Job evaluation is both a tool and a process: it provides a consistent framework for considering the relative value of jobs within the context of your organisation, given your mission, strategy, and current circumstances.
AI uses existing generic information to determine the value of a role. It uses titles and, perhaps, brief descriptors, but it is unlikely to customise for context. It also relies on unvalidated data. Even AI says human oversight is required to verify AI insights with expert advice and robust evidence to avoid bias. Because AI uses existing data, it can also replicate existing bias.
A more robust approach to determining pay levels includes undertaking a work value assessment as a first step, using a structured analytical approach such as job evaluation. Accessing a comprehensive and robust dataset, using the outcome of the job evaluation process, provides another strong foundation for your remuneration decisions. Using job evaluation to underpin remuneration decisions enables an in-depth understanding of the role, provides a greater degree of accuracy in the assessment, and ensures relevance and strategic alignment for your pay policy and outcomes.