The HRO Today Talent Acquisition Summit and the HRO Today CHRO South were held in North America in September, so I had ample opportunity to interact with top HR leaders. The good news: No one at either conference mentioned artificial intelligence. Of course, I am kidding. The conversation about AI is far outpacing the capabilities of the systems. That much is clear.
I’m a strong supporter of AI. It will fundamentally change how people interact with technology. Technology can help us become better informed, make better decisions, and work more productively. However, I don’t expect it to cure acne, make teenagers easier to live with, or prevent the common cold. In short, I believe in AI’s promise—not the hype. It is also important to remember that the AI companies we discuss today are primarily large language model providers. There are thousands of start-ups using these search tools, analytical tools, and agentic tools that are available to create custom applications. HR has been a popular market for these AI start-ups, and there has been considerable investment in these software offerings. However, most start-up technologists have never lived in the “applied” world of HR.
One of the core assumptions of the software, particularly the agentic software, is that it can replace human workers. While this may be true in some administrative and analytical functions, it requires knowledge transfer from those “pesky” humans that may not be so cooperative. As I have said many times in speeches, AI means different things to different people. When I say AI, your Board hears “dividends,” your executive leadership team hears “optimisation,” and your employees hear “layoffs.” As a result, implementation can be mired in process documentation and transfer and training of the bots and can take longer than expected and be far more costly.
We also have the phenomenon of unreliable or hallucinatory AI. These effects can be unanticipated. As the products get more sophisticated, this issue should be less troublesome, but unreliable performance can cause problems.
Lastly, you need to think about your providers. One example I often cite in speeches is about the rise of radio entertainment, another quantum leap in human communication. However, during the 1920s, about one out of 10 new radio stations went bankrupt as the market weeded out poor performers. We saw the same thing in the internet era. We all pretty much use Google, but does anyone remember AltaVista, Lycos, AskJeeves, or Webcrawler? I am not suggesting that you do not buy innovation that comes from start-ups and only use the big players, but I am saying that your procurement function should do proper due diligence on software providers and their financial resources and investors.
The prudent approach is to resist the AI craze fueled by punditry, headlines, and promises of massive savings. HR leaders who are satisfied with their AI progress consistently recommend first evaluating needs and identifying where AI can increase productivity, eliminate drudgery, and handle analytical tasks better suited to technology. Only then should you select the best solution from providers likely to withstand both the test of time and the inevitable shakeout in any emerging market.
That is the prudent approach to AI.
Elliot H. Clark
CEO
