How are you measuring and demonstrating the ROI of your AI investments at scale?
Sort by:
I am measuring my reduction in headcount (salaries/benefits) in Invoice to Cash (I2C), as I have gone from a seven-person department, including a manager, to two upskilled/promoted people. The savings helped fund the salary increases, AI for A/P software, and the AI for Expense Audits and Credit Card software we are currently implementing.
AI investments should be measured both hard impact and strategic value.
Hard impact where you can measure - no of hours saved, productivity increase, shorter recruitment cycle.
Strategic values can be like - Candidate experience, Hiring manager & recruiter experience.
It is important to start with understanding the business value you expect to achieve with your AI investment - Revenue Growth, Employee Retention, New Product/Service/Market etc. Once you have this, you can measure and demonstrate ROI against these KPIs.
In HR we are just embarking on the use of AI in our operations. So we are not yet at scale and are at a "crawl" stage. Interested and following..
There's a lot of good feedback on this. We had considered some of these and decided that it couldn't really be tracked as savings. The reason is because the technology is enabling things that wouldn't have been practical before. If it would have taken 2 FTE years to do analysis on something, we would have never invested that - we would have talked to the executive leaders and made a guesstimate on the analysis. So, if you use AI and it takes you 4 hours instead of 4,000 - you can't claim the 3,996 as savings. Being aware of the things that we are doing that were previously impractical, we use the term "Digital Yield". The "Digital Yield" is the number of hours it would've taken in the previous toolset to produce the same outcome. My team does a lot of things "off the record" to help other teams and I'm seeing 30:4 ratio on my teams investment and digital yield results. Four hours of "help" yields 3 weeks of "Digital Yield" on average. Some higher, some lower.