Legacy Meets Machine Learning

Spur of the moment decisions are a constant in our life. Chicken, Beef or Tofu? Take the highway or use the side streets? Coke or Pepsi? Use Google Maps or Based on our experience, we use such data to make immediate decisions.

Businesses can now make millions of intelligent, immediate decisions given the utilization of artificial intelligence and machine learning. In my 25 years of technology, Operations and eCommerce, the majority I are dependent on their legacy applications – in one way or another. Online up-sell/cross-sell techniques, personalized ads, banner placements and inventory messaging or a snippet of these immediate decisions made today.

What drives success for these immediate decisions to populate an ad or messaging? Timing.

What makes a retargeted ad successful? Context, location, device, behavior and weather determine if the ad will convert into a purchase, a click or any other trackable conversion.

With that said, consumers (B2B, B2C, B2B2C) are constantly changing how they interact by device, time of day, browser type and this is all changing over time. For businesses, especially those utilizing legacy platforms, it is imperative there is an accurate snapshot at a single point in time.

In most legacy enterprises, there is a plethora of data stored within enterprise firewalls and in data warehouses and lakes. There are also petabytes of new data from IoT devices, server logs, and media. Utilizing AI to make decisions is viable because the raw inputs exist. And it’s not just a nice-to-have. It’s crucial for success, especially for the legacy applications that have made the business successful in the first place.

Most businesses have lengthy Master Data workflow meetings, Data Intelligence meetings and the one question that needs to be answered is: What technology or software needs to be in place in order for the business to make spur-of-the-moment, intelligent decisions that drive value and revenue.

We have defined this into three elements:

  1. Accessibility: Ability to access the most recent data in an ML model so users to take action while it is still relevant to them.
  2. Constant Change: User behavior is constantly changing and ML models need to keep up with the change and therefore, the models need to adjust according to change
  3. Experimentation: ML models and, more importantly, the data scientists that are driving such models need the ability and business support to experiment with new features and new data models in order to push the envelope and drive success.

Leave a Comment

Your email address will not be published. Required fields are marked *