Reinventing technical training to support the complete learning process

Reinventing technical training to support the complete learning process

Reinventing technical training to support the complete learning processReinventing technical training to support the complete learning process
Analytics Maturity Curve

Is Your Organization Ready for AI?

Not So Fast - Intelligence First

There is no doubt that the catch-all buzzword of our time is AI - or Artificial Intelligence.  This is a weighty concept with all sorts of implications, and there are plenty of sources of information available on the topic.  The application of AI in the world of Analytics is one of the most compelling - mostly because we are not talking about some futuristic concept - we have been using AI with data for decades.  Let me explain...

As you can see in the above graphic, at the far right, is a representation of our classic reference architecture for data and analytics (source data, EDW, etc.) - and there on the very top sits AI.  Note that just below there are the building blocks of AI, namely: Machine and Deep Learning (ML and DL) Frameworks and Data Lakes.  Think of Data Lakes as the holding pens of all possible data assets that we might find useful in an AI application.  Think of ML and DL as the algorithms that are used to refine that data via sophisticated models which ultimately generate predictions about future outcomes.  Feeding those predictions into user-accessible applications, or into other machine processes is the "endpoint" of an AI application.   OK, great, so let's do it!

Not so fast - notice the position on the Y-axis (scale and maturity).  The graphic is meant to reinforce the laws of scale and maturity, meaning you have to get the OTHER stuff right FIRST.  IBM is using another term that helps remind us of this reality "The AI Ladder".  Tools as mature as SPSS have been enabling this type of advanced analytics for a very long time.  So, it's important to remember that ML and AI are NOT new, however they have become much more accessible, and their value has become much more well-understood.  However, leaping into these capabilities before your organization is ready carries a tremendous amount of risk.  The time-tested principle of garbage-in, garbage-out once again rears its ugly head.  

Intelligence First is our guiding philosophy that forces a sober, honest, and thorough assessment of where an organization sits on the maturity curve.  Building expensive predictive solutions on siloed and/or soiled data will not add value to the business.  Those initiatives will divert resources away from activities that are needed to shore up your analytics foundation.  Let's work together to develop an Intelligence-first analytics strategy that will enable you to deliver on all rungs of the AI Ladder!

The final, and probably most significant success factor seems too obvious, but is frequently overlooked: Have a viable business case with measurable ROI before embarking on AI projects!  This is another of those dirty little secrets in our business.  From the early days of information management, the FOMO (Fear Of Missing Out) associated with being late to the party, whatever the craze of the time happens to be - ERP, EDW, BI, Big Data, Cloud, AI - overpowers our common sense and organizations rush to implement, skipping key tasks such as cost-benefit or ROI analysis, strategic alignment, and even basic requirements gathering.  Yes, there is risk to falling behind the competition by adopting game-changing technology too late, but don't ignore the risk that rushing into such ventures brings with it.  We are still managing bungled BI deployments that made similar mistakes 10, 20 years ago.  All too often issues are not caused by poor technology choices (although that is a convenient scapegoat), but rather because organizations failed to align the tech with strategic objectives.

Let's do this thing

Regardless of your technology choices, if you plan to climb the AI Ladder, partner with the right team to build solid footholds throughout the journey.