Augmented Analytics – The Next Evolution in AI

April 22, 2019

I recently wrote about augmented analytics for a Forbes Technology Council column, after reading an interesting article on the subject in The Wall Street Journal. As the news article mentioned, while AI is one of the most important technology advances of our era, it still has a ways to go before it can identify cause and effect relationships. To do this requires massive data-sets, that will be more readily available through augmented analytics, or AI-as-a-Service. Once it accomplishes this feat, however, it’s set to bring AI to the masses and seriously up the game for what AI can accomplish.

According to The Wall Street Journal article, determining causal relationships requires the ability to sift through tons of “large and noisy data-sets to detect faint signals, or the proverbial needle in a haystack.” It’s one thing to use AI to find patterns that can indicate growing traffic issues in a community, for example, and quite another to determine what is causing the back-ups. That level of insight needs massive and more granular amounts of data, and that’s where augmented analytics comes into play.

As the article explains augmented analytics will be delivered via pre-defined AI as a service, or pre-populated data that can supplement the limited data-sets data scientists may already have.

Earlier this year, Gartner identified this as an emerging role for 2019 – The Augmented Developer, who will build advanced algorithms based on augmented analytics. Gartner also identified Augmented Analytics as a Top Ten Strategic Tech Trend for 2019.

Yet, I mentioned in the Forbes article that while it may be considered a top tech trend for this year, the notion of AI as a Service, providing augmented data and analytics, won’t be realized for some time – perhaps not for at least three years or more, since there are key business challenges that must be overcome. There needs to be greater domain knowledge in the industry in order to create pre-defined data-sets; and then there’s the challenge of enabling generic data to meet unique business needs.

While the lure of working with providers that claim to already offer pre-defined AI as Service is enticing, buyers need to be informed. They need to ask the hard-hitting questions, such as:

  • Do you have domain knowledge? How much do you know about my specific industry to provide me with the data to address my business problem?
  • How easily can I integrate my own data with the pre-defined data-sets?
  • How can I measure the algorithm’s performance?
  • Will you remain a long-term partner that can work with me to continuously augment and train my data-sets?

Once pre-defined AI as a Service takes hold it will change the role of AI as we know it, but it’s going to take some time. In the meantime, customized AI solutions are available today that are bringing the level of personalization and unique business understanding to help you solve key business challenges. There’s really no need to wait for what comes next.

 

Leave a Reply

  • (will not be published)