AI Advantage

WHERE ARE YOU ON THE ENTERPRISE AI CONTINUUM?

One of the things that makes the current AI landscape so complex and confusing is the term AI itself. Right now, it may seem like all AI is generative AI (ChatGPT, Bard, etc.) but AI actually comes in lots of different flavors and each one has different attributes. We call this spectrum the Enterprise AI Continuum and we’ve created a handy graphic and companion article to guide you through the different forms of AI—from the least sophisticated to the most advanced (so advanced, it doesn’t even exist yet)—so you can better understand how to use each and every form of AI to your advantage.

Roll over and click on each area of the graphic for more information. Scroll to read the full article.

Starting on the far-left side of the continuum is Automation & Analytics. These are capabilities such as dashboards and data visualizations, advanced or interactive reporting, and robotic process automation. There is a lot that can be learned and unlocked from automation and analytics – especially when you align it to your organizational decision-making needs – but it’s only the very beginning of the continuum of AI.

The next step in the continuum, in terms of degrees and levels of sophistication, is Data Science. Right now, there is a fair amount of marketplace confusion around the differences between AI, ML, data science, and generative AI. At its heart, data science is math. It’s math models that help you decompose and solve business problems using common math tools, formulas, and algorithms, so you can predict behaviors. These can be market behaviors in the form of demand forecasting so you can reduce risk and maximize opportunities. Or how customers are likely to behave on a company’s website and in its retail stores so you can have the right product in stock and offer customized promotions. It can also be the behavior of your assets in a manufacturing plant, predicting equipment failures and maintenance needs so you can reduce service downtime and operational costs.

You can think about the first two stages in the Enterprise AI Continuum this way:

• Automation & Analytics tells you what happened or what is happening
• Data Science tells you why it’s happening, what’s likely to happen next, and what to do about it

After Data Science comes Machine Learning (ML). ML is also based on a set of math problems; the big difference between the two (because a lot of times they answer the same business question) is that machine learning is a little less transparent. It doesn’t tell you why it’s giving you the answer that it does. Here’s an example: In working with a large hotel chain on a segmentation initiative for their properties, we considered two different classification techniques: one based on traditional data science methods and the other based on machine learning. The data science technique will make it very clear why it puts each hotel in the cluster that it puts them in. The machine learning technique, however, will not tell you why; it just tells you to which cluster each property belongs. You’ll get generally the same answer regardless of which technique you use, the big difference is that one tends to be more transparent than another. The downside of the more transparent approach (data science) is that it requires more handholding – supervised learning as opposed to unsupervised learning.

To the right of Machine Learning is the hot topic du jour: Generative Artificial Intelligence (GenAI). GenAI is what’s dominating the AI conversation right now, hitting the market in tools like ChatGPT, Bard, and many others.

What differentiates GenAI from the AI that has come before it is that it creates new content. You ask it a question and it generates an answer to that question. But it’s not generating content out of its own “brain.” This is not like the movies where the AI takes over the world. GenAI doesn’t have a brain. It’s run by specific models that are trained on incredibly massive data sets. It’s actually generating new content by reconstructing existing information or images that it has already “seen” before.

Everything from number 3/Machine Learning to the left is almost always within your firewall or within your cloud tenant, so nothing is ever escaping your environment. It’s when you start thinking about how to leverage GenAI that security becomes a really important conversation to have because you are potentially interacting outside of your firewall. More on GenAI implementation patterns and the security considerations of each to come.

It’s important to note that there is another type of artificial intelligence: Artificial General Intelligence (AGI). And it is exactly what you think of from sci-fi/dystopian future movies, where the AI can think and reason like humans (and in film, usually does take over the world). But you’ll notice there’s an asterisk on AGI in the graphics because, luckily, this technology doesn’t exist yet. So it’s one less thing you need a rapid deployment strategy for this year!

Wondering how to leverage the Enterprise AI Continuum to your advantage? Reach out to us to discuss how best to align your use of AI – in all its flavors – to your organization’s goals so you can make investments that truly pay off for your business.