Systems | Development | Analytics | API | Testing

The hype around AI is deafening, creating a noisy and confused environment. Organizations are looking for clarity on the role and application of AI in bringing more value to data, and for help avoiding the pitfalls that surround so many of these projects. These pitfalls include error-prone decision making and unintended consequences based on flawed data or “black box” algorithms, and gimmicky approaches including simple search and bolt-on AI that overpromise and under deliver.

My first car had a manual choke. It would change the ratio of air and fuel to allow the engine to start - you would pull it when starting the engine cold, wait a bit and then set it back. The car had many other peculiarities - a manual transmission (of course), no power steering, and the engine had a tendency to overheat in the summer - or, if going uphill, year-round.

Not so long ago, our CFO was looking at changing our financial system and it reminded me of a similar situation that happened when I worked at National Australia Bank many years ago. We changed our entire ERP because the organization needed to slice and dice their data differently. What they could have done instead was use analytics differently to extend the life of the products they had, so that’s what we’ve done.