Where to apply AI First

There are several ways to solve business problems and seize new opportunities using AI. The key for most businesses is to answer the question "what problem to solve first?" and "how to solve it?". In other words, what should be your AI Strategy? (the "what") and what capabilities you need? (the "how"). With that premise in mind, I want to list a few approaches to decide the "what problem to solve first?".

Data Intensity

Determine the data intensity of your operations and apply AI to the most intense operations first. Data Intensity is a result of how your business is designed. If your business needs a lot of people or has a wide product mix or perhaps is an international business with several related enterprises, then your business is complex and will have higher data intensity. Within your business, find the function or the supporting divisions that are most intense. Apply AI solutions there first.

Customer Driven Ranking

Make a list of nearest, second nearest and third nearest processes from the customer and subjectively rank those processes by importance. This process usually results in application of AI in the CX (Customer Experience) Value Chain, the Fulfillment Value Chain and the Production Value Chain. This disfavors business support functions such as IT, Finance, HR for example. If your organization is hyper-focused on customers this is a great place to start.

Business Performance Improvement

In your financial statements, in your BI reports, find those items that are hardest to grow or hardest to reduce. Apply AI solutions in the functions that have those items. This is the simplest to grasp as it has the "follow the money" vibe. The problem here is that not all big numbers are caused by issues that AI can solve. If you are a jeweler who purchases precious stones, this method will mislead you into trying to apply AI to optimize purchase prices or vendor relations while you might benefit more from customer experience improvements. On the other hand, if you are a logistics company this method may lead you to making better maintenance decisions on long haul equipment for example.

Simple First

If you are truly starting from scratch, this is a common-sense approach to get started. Ask for project ideas from your function managers ranked by their subjective evaluation of simplicity. Pick one and implement all aspects of the AI Platform for that project. Then, pick the next one. No management science here!

These are just a few approaches. In reality, Enterprises can be messy and chaotic. Out of the mess and chaos, good things do happen. You will need to be the leader who can see through the fog and lead your enterprise to the other side.