Training Watson Video+Blog by Joanne Wright, IBM's Vice President, Supply Chain
Print this Article | Send to Colleague
You've heard it time and time again: we're in the midst of a digital revolution that's changing the way we do business. Today's supply chains are straining under the weight of mountains of data generated by the Internet of Things (IoT), mobile, social and other digital technologies. The opportunity for insight is greater than ever, but true visibility still seems frustratingly out of reach.
Artificial Intelligence (AI) is the anecdote to digital disruption, extracting new, actionable insights from data at astonishing speeds. Already, leaders are looking for the best way to capitalize on AI to drive competitive advantage.
But what few are prepared for, is that AI requires training in order to deliver the outcome the business requires. It's not that hard to whip up AI code that employs learning techniques -- what is differentiating is the body of knowledge that you provide and having your supply practitioners invest the time to train your AI solution of choice on how to use your data.
Let me give you an example. IBM started its own cognitive supply chain journey with Watson three years ago. We knew that to maximize the potential of AI for our supply chain, we needed to be prepared to train Watson. As my team leveraged Watson, in turn it provided the fuel for it to learn more and get better at augmenting my team's skills and ability to resolve disruptions.
Watson: In training
We used subject matter experts to teach Watson our processes and playbooks for disruption mitigation. Our supply chain practitioners developed unique question-and-answer pairs to feed Watson, and tested it to ensure it answered with the same degree of accuracy as our experts. We helped Watson continue to learn by providing it with all the possible variations of answers to our most critical supply chain questions. We continued this process iteratively until we were confident that Watson could answer all our questions. This helped Watson understand the intention behind our questions, helping it get to the root of our problems quicker.
Reaping the AI rewards
Now, when one of our supply chain practitioners is working on a part shortage, he or she has immediate access to inventory data, alternate parts, real-time supply, in-transit data and supplier information. Watson is always scanning ahead, anticipating potential problems, and providing possible solutions. Also, data visualization, coupled with rapid data retrieval, provides insights that we otherwise wouldn't have, and much quicker than traditional ERP systems will allow. My teams get an aggregate view of supply chain data without having to access multiple systems, submit complex queries and run multiple reports. The result? My team is now free to tackle problem resolution straight away.
Moving forward, we'll provide Watson with even more internal data and incorporate external data -- like weather, news and market trends -- to give us even better end-to-end, supply chain visibility.
Are you ready to learn more about what it takes to train Watson for your supply chain? Read The AI Journey: Artificial Intelligence and the Supply Chain Smartpaper.
