Conversational AI has been in the headlines recently. In the last few weeks, we have seen many discussions around whether conversational AI programs have come close to being sentient. Another key area of discussion that I believe we are not having and is equally, if not more important, is how to extrapolate these advances into fields like operations and supply chain. Supply chain planning is one area where this technology can play a critical role, specifically when we envision autonomous supply chain planning. SAPinsider will be publishing a couple of reports focused on supply chain and inventory planning this year. While our July 2022 report, "Supply Chain Planning in The Cloud" covers all elements of supply chain planning, our November report, "Inventory Planning, and Optimization" focuses specifically on inventory management. This article discusses some of the ways conversational AI can be used for inventory planning and management.

What is Conversational AI?

In very simple terms, conversational AI algorithms are AI-enabled programs that can interact with humans in a meaningful way, deciphering human language ( voice or text) to formulate an appropriate response. We frequently encounter an elementary version of text-based chatbots on websites these days. Alexa and Siri are examples of voice-based conversational AI.

Applications in Inventory Planning and Management

There are two ways conversational AI can be leveraged for Inventory planning and management. One way is as a standalone capability whereas the other is where it works in tandem with other mature emerging technologies. The following applications cover both these areas: Augmented Inventory Analytics The conversational AI model used in this category is Natural Language(NL) database interfaces in technical jargon. These models translate natural language into database queries to extract the data desired. This is rapidly becoming a mature technology with more and more solution providers embedding this capability into their BI and analytics solutions. Examples in this category can be interacting through text or voice with a BI tool with questions like:

• What is the current on-hand Inventory for SKU ABC?
• How many days of inventory does that translate into?
• When was the last order placed? What was the quantity?
• How many units are scheduled to be delivered in the next 30 days?

Analyst Tip: If your current BI or analytics solution does not have this capability, there are solutions available like SAP Conversational AI that you can use to build a natural language database interface. Then there are always options to use open-source technologies. Unlike the example questions, if you leverage open-source technologies to build your own “augmented analytics” capabilities, you can create scenario questions. An example is: Will we run out of inventory in the next 30 days? With dialogue planning models, the solution can leverage the answer to questions listed above to answer your scenario question on stock-out. With other mature and emerging technologies In this case, conversational AI can be integrated with real-time visibility data sources to accommodate aspects like pipeline inventory. An example is: What is the current status of the latest shipment from Shenzhen, China? These solutions can be embedded in Industry 4.0 to enhance its capabilities. Using their mobility devices, supervisors can inquire about the status of an asset or the overall process. I can foresee some possible integrations with blockchain technology as well. Developments in machine learning, specifically deep learning, transformers like BERT, and reinforcement learning, have provided us with more tools to build more advanced and customized conversational AI tools, designed around our unique inventory management and planning challenges. Analyst Tip: When building advanced conversational AI capabilities, explore possibilities of "interaction" with other business applications. While NL database interfaces are pretty common now, there are other possibilities, where you can also use chatbots to execute applications and algorithms. Then there is the opportunity to leverage data being generated from conversational AI for analytics. If you are leveraging a platform like SAP BTP, you can use the platform as a bridge between conversational AI and other applications, like SAP Analytics cloud.