No Predictions 2018—Thriving on Data

Publish date:

My AI top tip: focus particularly on deep learning and neural networks, as most of the recent innovation breakthroughs are thanks to them.

One does not exactly need psychic capabilities—although they would still come in handy for Psychic Pizza delivery—to see that data takes center stage in the 2018 business theater.

And this is not necessarily a big deal anymore either, as the Volume, Variety and Velocity (and any other “V” you would like to add) of Big Data are now commonly available through all major BI and analytics platforms. So maybe we should simply talk BI and Analytics again next year (or DFKABD—Data Formerly Known As Big Data) and just leverage all the goodness, plus cloud, self-service, advanced visualization and DevOps to make anything data shiny and new.

Of course, you could join the AI Washing movement—and rebrand all that is even remotely associated with decision support, business intelligence and analytics as Artificial Intelligence. That would leverage the incredible amount of current enthusiasm for AI in even the most conservative of boardrooms, but the more inflated expectations are, the louder the bang if expectations are not met.

Obviously, AI has transformative, disruptive power. It can cut out intermediates, render human tasks obsolete, and even go into highly complex, holistic areas that our own brains can no longer deal with. With that much Black Swan potential, ethical, political, cultural and society-wide discussions need to take place. And this will take time. It shouldn’t keep us in the meantime from getting hands-on with AI right now, to reap some early benefits (our latest AI report contains dozens of suggestions) and to start mastering key technologies and capabilities.

My AI top tip: focus particularly on deep learning and neural networks, as most of the recent innovation breakthroughs are thanks to them. Also, beware of getting trapped in the lower levels of the AI solution stack. Sure, NVIDIA and IBM produce high-performance, dedicated infrastructure components, and open source-based platforms such as TensorFlow and MXNet are the crucial foundation (watch ONNX for an evolving open standard). But higher-level libraries such as Keras and Gluon are even easier to use and more productive for a larger group of developers. Furthermore, many already trained and configured AI solutions are available—both through APIs and as-a-service applications—so you don’t have to science yourself all the way out of it on your own.

Then, finally, never waste a good crisis next year. Data privacy and security are top of mind for many organizations, both inside and outside Europe. If you’re not aware yet, by all means do a little web search for GDPR, if you happen to have a few hours off. Trust has a balance that you can add to or subtract from. The organizations that do not consider data regulations as a valley of doom—great risk that needs to be mitigated—but instead as a way of boosting trustworthiness with customers and getting a better grip on data, will truly thrive.

I wish you a 2018 in which you raise your corporate IQ considerably, taking a hands-on, applied approach to delivering with data and digital. By the way, did anybody order a pizza?

See the full No Predictions 2018 overview here

Related Posts