Practical Approaches to AI: Don’t shoot for moon

By Molly Chakraborty

“A dwarf shooting for the moon” – is a phrase in my native language. In regards to AI or, for that matter, any new emerging technology we tend to look for innovative ways to solve our difficult problems. AI has been no exception. The concept of AI using neural networks is not new. It has been the topic of Ph.D. thesis for the last 20 years. However, are we shooting for the moon on AI? Very few CEO’s can take an emerging technology, make a product like iPhone or FB out of it, and reinvent the concept of consumerism. In fact, the same CEO’s may not be successful in their second act for similar endeavors.

Should we use IBM’s Watson project to make cancer prognosis automated or FB’s approach of making robots learn and talk in a natural language? Those are optimistic mega endeavors on AI that will change the way humans live on earth. However, at the same time, we can take baby steps in areas that the market can easily chew and digest.

Following are some areas where we have seen AI has been perceived as promising and delivering without huge investments: Improve Process, Insights into complex systems, and engaging with people. Examples of process automation in the supply chain are processing a back-to-back order reading a customer PO and creating a sales order, performing pick release, ship, and invoice, integrate AP/AR, perform three ways match to post in GL using bots. The key to this approach is the ability to work across multiple end-to-end systems. The thumb rule is if you can outsource a business process, you can automate. Cognitive Insight examples in the supply chain are monitoring failure points of machines by tracking and interpreting sensor data, planning for spare parts and scheduling service or predicting customer intent to buy. Deep learning, probabilistic matches, speech recognition, image recognition play an important role in this application of AI. This area, in particular, does not threaten any workforce, as today this kind of capabilities do not exist in any streamlined way.

Performing these tasks need highly intellectual and skilled workforce, building and sustaining which becomes a difficult task for any organization. Cognitive engagements in the supply chain could be taking orders from customers, answering their questions about availability, shipments; collaborating with supplier and partners; holding sales and operations meetings to present dashboards and come up with consensus forecast supporting the entire session in multiple languages. With time, cognitive engagements will change how companies do business.

Digital disruptions are here to stay. We need to make prudent decisions how we merge the technology with possibilities without signing up for major failure points or destabilizing the workforce and use it to our benefits. Understanding the technology, identifying pilot projects that are suitable to take the organization forward, scaling up and knowing ROI are key factors to make such projects and endeavors a long-term success story.

Based on article “Artificial Intelligence for the real world” by Thomas H. Davenport and Rajeev Ronanki
Harvard Business Review, Jan – Feb 2018.


About the Author

Molly Chakraborty is co-founder and managing partner of Trinamix Inc. Earlier in her career, Molly worked with Oracle as a thought leader in technology and application of Oracle products. Molly is a gold medalist from Jadavpur University. She has demonstrated leadership in successfully enabling major SCM and Manufacturing initiatives while taking Trinamix from a small startup to a globally recognized system integrator in the process.


For more Information Contact Us