Introducing work intelligence systems
Over the past few months, my colleague Matt Mullen and I have been taking briefings, researching academic and industry sources, and brainstorming a new area we call “work intelligence.” It’ s not something we have invented: It’ s a movement we have watched emerge during the past couple of years. In simple terms, we are talking about using AI to converge data sources and insights from employees, customers, operations, and processes into a unified whole, the goal being to drive better business decision making. If you are as geeky as we are, it’s fascinating stuff. Not only is there an emerging market, it will likely be a big one. But underlying all our research and discussions has been the uncomfortable reality that, although work intelligence makes perfect sense—the technology is available today to make it work—it doesn’t fundamentally work. Or, to be more specific, it works, but not as well as its champions would have you believe.
Shortcomings in KM problem-solving largely due to a mismatch between what the technology promises and what it can actually deliver are recurring themes in this column, but 20 years ago, in the dot-com era, the problem wasn't a lack of commitment or ambition; the technology available then just wasn't very good. Fast-forward to 2023, and the technology available is mind-blowingly good, reliable, and, in most cases, affordable and accessible. The problem today is not the shortcomings of the technology. Instead, it is the lack of a consistent approach to using it, along with poor quality data, and the belief that more data and processing power will resolve any kinks in the system.
Starting out
Let's start with a consistent approach. If you haven’t set clear business goals and defined a short-, mid-, and long-term path to achieving those goals, the odds are that you will fail. Then there is the issue of data. Technology, particularly AI, needs clean and accurate data, but seldom gets fed that degree of data quality. And here's the critical point: It never will. Trust me, at the start of every project, the need to clean and maintain data is stressed, but it's a message businesses don't want to hear. Think about it this way: Data from an IoT device is probably pretty darn accurate. It will read the temperature, location, humidity, and if a device is switched on or off. It won’t be perfect, but it’s probably good enough. But compare IoT data with the millions of documents stored in SharePoint or an HR system. The quality, accuracy, and consistency of the data stored there will likley to be comparatively woeful.
Some tools (typically AI-based) can improve data quality, but they will never be 100% accurate and always lack human context. But that doesn’ t mean you shouldn’t use AI or a new-fangled work intelligence system in the future. Far from it. Technological advances are significant and can bring huge benefits, but only as long as you understand that they can advise, augment, and support, but not replace, you. Even the most advanced AI systems, fed with the highest-quality data, are surprisingly limited in what they can do. They do some things exceptionally well, better than any human. But when it comes to running your business, making critical decisions, planning, and transforming, they should always be in a supporting, not the leading, role.