In our last post we started reviewing 20 analytics predictions being peddled by the major analytics futurists and analytics sites. Why? Because while overly optimistic futurist authors rarely get it right, their predictions do point out two things. What should be done — and isn’t getting done — and where the space needs to go.
And even though 19 of these anti-predictions won’t (fully) come to pass this year, we started reviewing them one by one to give you a reality and indicate what is likely coming sooner than later, and what is still a pipe dream. Most of the predictions we reviewed yesterday were those that fell into the “aren’t happening” or “aren’t really happening at all” (because they are more of the same old, same old) buckets, but today we get to some that will start to materialize and the one, yes one, that is 100% true — and that you need to be fully aware of.
So settle in and let’s finish this.
AI becomes more mainstream
Well, acceptance of AI will continue to become more mainstream, but considering that most “AI” providers are actually providers of “Artificial Indirection” and have no AI at all, not even at the level of “Assisted Intelligence”. Most providers of “AI” are just providers of RPA (robotic process automation) at-best, and a configurable rules-engine at worst.
A few vendors are offering multi-hybrid analytics solutions, and a few more will, but there will be nothing new. It will be one solution for integrated in-platform analytics, another for do-it-yourself analytics, and possibly an in-house developed third for database management and cube construction. But there’s going to be no significant changes here — most practitioners are going to use what their vendors give them.
Analytics will become usable by business analysts
Well, this one is half true. With recent advances in user interfaces and usability, it will become more usable … but … only to the better half of the business analysts … and … only with training. And this is where this particular prediction fails. Training has been high on the priority list for a decade, and it’s also been high on the “cut when budgets need trimming” list for a decade as well. There will be little to no training as per the norm, so only the most dedicated will self-learn and use it.
Data governance takes centre stage
This prediction is likely to come sooner than you might think, but not in 2020. Until there is a big cost associated with the lack of data governance, like training, it’s going to remain high on the priority list but not going to get centre stage. This will only change when lack of governance risks a huge fine or a large organization loses a major court case with a large judgement that was the result of lack of governance (which resulted in data exposure) which could have happened to any governance.
AI ethics standards will emerge
We all wish this will happen, but as with data governance, until a large organization loses a discriminatory court case as a result of an AI decision, and the court holds the organization responsible for that AI decision, no one is going to put any real effort, beyond lip service, into AI ethics. At least from a vendor perspective. A few lawyers hungry to make a name for themselves might, but that’s about it.
Analytics will hit the C-Suite
Re-set the woody woodpecker laugh track. If the average business analyst is not going to get much more involved with analytics, then you can bet the average C-Suite executive is not going to get much more involved either. They might get better reports and dashboards, but that’s it.
Intelligent assistants that connect the dots will become more pervasive
This is another half-truth. “Intelligent assistants” that allow a user to interact with the application in natural language, and especially English, will continue to infiltrate S2P platforms, but as to connecting-the-dots … not likely. That will require true embedded machine learning technology, and that’s still far away for the average provider.
Open source is going down the drain thanks to cloud platforms
This is yet another half-truth. While it is true that as more and more providers lock into a cloud platform (such as Azure, Microsoft, and Google) they will lock into whatever analytics are provided in the platform, this is not going to stop open source efforts — although uptake may trickle off for a while.
Effective implementation will continue to be a challenge
This is mostly true. Effective implementation will continue to be a challenge for the majority of organizations, and only a few best-of-breed providers will see the challenge of effective implementations decrease. As data continues to proliferate, especially considering the average quality of data, analytics will continue to get more challenging on the whole.
And now, finally, the one prediction the doctor found that is 100% accurate.
Companies will continue failing analytics & AI transformations
This is absolutely true. Considering that analytics requires good data and AI requires lots of good data, good algorithms, and experts to guide the algorithms, and most companies have poor data, poorer algorithms, and a dearth of experts … and often rely on vendors who peddle applied indirection, the doctor expects a big uptick in failures until the space educates themselves on what AI truly is, what the levels are, what is actually out there, and who is actually offering it.
For details on what the levels are, and what is coming, keep your eyes on SI and SM, and if your organization has been smart enough to subscribe, check out the doctor‘s pieces over on Spend Matters Pro on AI in Supplier Discovery, Sourcing, Optimization, Procurement, and Supplier Management.