The artificial intelligence (AI) road from unlimited (yet largely nonspecific) potential to concrete, specific business benefits, is like taking a long road trip with kids — palpable excitement alternated with restless tension and cries of “Are we there yet?”
So, are we “there” yet? Well, no, but we are certainly closer than we have been, and by further examining the data that underlies these systems, we can progress closer to "there" by recognizing measurable ROI from the ability to make better decisions with AI-powered analytics technologies.
We’re at that inflection point in the AI hype cycle. Most companies that say they are using AI have yet to gain any value from their investment, according to the 2019 Artificial Intelligence Global Executive Study and Research Report from MIT Sloan Management Review and Boston Consulting Group (BCG). They continue to plug away, however, even though the payoff — new products, increased revenues and optimized efficiencies — is likely further out than previously imagined.
I’ve outlined the tremendous agility and precision that AI is poised to bring to the supply chain — as well as the importance of building in domain-specific intelligence that maps industry and function-specific capabilities with data to solve business problems.
Big Data That’s Getting Bigger
Just as Waze and other “smart” GPS systems use data to optimize the family road trip, I believe the road to AI returns will also be built on data. There’s clearly no shortage of it: The concept of big data — the large volume of structured and unstructured data collected by businesses on a regular basis — has been around for nearly 15 years, and in that time, it’s only continued to get bigger.
This data is coming from a variety of places — internally, it’s business transactions, back-office information, customer and prospect data, IoT sensors providing machine information, etc. The supply chain is a particularly rich source of data — each hop from raw materials through shipping to the end user provides valuable historical information. The real power comes from matching internal and external data with third-party data.
Companies realize that their operational data leads to valuable insights. Yet this internal information is biased and has gaps. So, organizations are increasingly looking to add additional sources of data to their analysis — anything from weather and demographic information to satellite information. While transactional data remains a foundational data asset, a 2018 Gartner survey showed that nearly half of organizations are using external sources. The most common of these are weather data — for example, by correlating sales with a weather stream, retailers can project demand for snow shovels.
The combination of enterprise data and public data powers better decisions — through “smart data.” There are a number of definitions for smart data. Dun & Bradstreet’s chief data scientist, Anthony Scriffignano, called it "the subset of that data that will actually apply to your problem — that can be used intelligently in a way that takes you toward a solution.”
While big data would be a long list of numbers representing weekly sales figures, smart data would be information that has identified the peaks and valleys in those numbers, leading users to make decisions related to inventory, logistics or pricing.
Making Big Data Smarter With AI
Humans simply do not have the capacity to analyze these large volumes of big data — there just aren’t enough data scientists to make it happen. Smart data, however, is a different story: It’s the difference between a list of numbers and truly actionable insights, recommendations and, as users gain trust, automated actions that ultimately drive improved efficiencies, optimized operations, increased revenues and decreased costs.
For the supply chain, AI provides tremendous potential to convert big data — RFID and GPS location data, point-of-sale transactional data and third-party information related to weather and traffic — to smart data. Think about a container’s journey from Asia to the midwestern U.S. It needs to be in Chicago by Sunday and Indianapolis by Monday, and if all goes according to plan, it should arrive in the port of Los Angeles by Tuesday — at which point, it will head west via rail.
Of course, things rarely go according to plan — Tuesday in Los Angeles becomes Friday, but the goods are still due in Indianapolis on Monday because the supply chain waits for no one and the end customer doesn’t care about port delays.
There is plenty of data to help drive a plan of action to mitigate this delay. In this case, big data is collected from internal sources — such as marine terminals, ocean carriers, railroads and customs clearances — to track the container’s journey. Relevant external data includes historical numbers related to port delays and rail and road arrival performance, along with projections for weather and even street traffic.
The default on-land carrier is a Class I railroad to Chicago; but the AI-powered analytics, converting the big data to smart data, recommends a better option, a choice between two road carriers that both project a greater likelihood of getting the cargo to Indianapolis on time.
You can repeat this example across other industries, using AI-generated smart data to power recommendation engines for e-commerce and media (which have improved dramatically since their early days, yet still have continued room for growth), new approaches to battling the flu or bringing a new generation of electric vehicles to market.
AI doesn’t just make data and technology smarter; it has the power to make businesses, processes and, ultimately, people smarter, too. We’re still in the early days for returns on AI investment, but through the potential of smart data, the business value it delivers will be worth the wait.
