“OK Google, what caused my project to be 6 weeks late and $300,000 over budget?”

While we aren’t quite there yet, the day is coming where we can expect a logical and accurate answer. 

A few years back, questions started appearing on maintenance and reliability forums such as “The outboard bearing failed on my vertical lift pump—does anyone know what caused it?”, and “What causes cracks in the attachment welds of induced draft centrifugal fans?”  Having been on the hook to analyze and solve problems like this during my prior days in industry, the simple notion of asking these questions was……well, naïve at best.    

When we first started seeing these questions, I’m sure most of my gray-haired colleagues laughed and shook their heads, just like me. Knowing that the questions came from a generation who grew up obtaining answers with just a couple keystrokes, it was to be expected. Interestingly though, several years later, I’m realizing that I may need to rethink my position because things are changing fast.

Through ongoing development of artificial intelligence(AI)/cognitive computing, the answers to complex problems are becoming more available (and getting more accurate) as data sets expand. Google foresaw this 10+ years ago. The underlying challenge is not necessarily providing answers/solutions to complex problems but amassing the data needed by the AI algorithms to draw its conclusions.  Is it a mystery why Google produced so many valuable apps and made them available for free? When you checked the ‘Terms’ box, whether you knew it or not, you started giving them all your data! That was your ‘payment’ for their ‘free’ apps.

Early prototype work done by IBM’s cognitive computing platform, Watson, has shown that impressive results are possible when cognitive computing is fed enough data. Google and Microsoft are working on similar services and making them available to developers. Regardless of the platform, as the data set increases, so does the accuracy of the answer. So, while at present I haven’t seen any application that can accurately answer ‘Why is my project 6 weeks late?’, the day is coming where those answers will be possible once sufficient data is available. It is merely a matter of time. 

If you would like to take advantage of these capabilities down the road, you would be well served now to take the following steps---consolidate your data, and structure it to be more consistent. The future benefit will be better, faster decision support, and it will move you into the proactive world of prevention that you are always thinking about, but struggle to attain. A recommended action is to move your desktop applications, spreadsheets and autonomous files into Enterprise applications where possible. Data in Word and Excel files stored on individual drives is very difficult to corral and subsequently won’t provide much value in the future.

Aside from the numerous benefits of having all data stored in a common location, data in Enterprise systems will yield higher quality results because of the consistent data structure and tagging capabilities used for reporting. Specific to root cause analysis, Enterprise applications already identify ‘common causes’, a key source of proactive solutions. With just a little more time, I think they will be able to provide probabilistic models that tell us what, and where, the next failure will likely be.

Exciting stuff! Who knows? Maybe someday cognitive computing will predict our next president!  

Learn more about Sologic RCA