August 20, 2018
6 min read

3 Ways to Apply AI for Real-Time Retail: Part 2

AI has become a well-discussed term in recent years, finally moving from sci-fi pipe dream to real-deal opportunity for businesses to improve efficiency and service. As everyday shoppers, we experience the effect of AI as recipients of personalized advertising and product recommendations. That’s not all it can do, though, as AI is becoming a powerful tool for in-store retailers to turn imprecise and inefficient backend processes into value-drivers.

But why do we need AI for this? It’s not like retailers have been starved for data analysis or don’t know how to run their businesses based on decades of experience and customer relationship building. And yes, technology has been improving nearly every aspect of retail for years. But even when retailers adapted advanced rules-based algorithms to ease the burden, their improvements were capped by the limited power of historical data to tell the full story – which it really never does.

In the last blog, we explored how replenishment can be automated to optimize sales with AI. Now, let’s take a look at two ways that AI can impact marketing processes in ways no other technology quite can.

Digital Signage

For well over a decade, retailers have relied on digital signage in stores to promote dynamic messages, ranging from store alerts to promotions. In some cases, it’s not only a driver for sales, but also a space that can be sold to CPG manufacturers as a branded display at the point of decision.

Yet, many retailers have failed to succeed in using these signs because they weren’t ready to hire people to keep the program effective – it takes a lot of work to create and update the content. It’s also really challenging to dictate behavior with a message or change mass buying behavior this way unless there’s a great discount available.

But now with AI, retailers can predict product sales far into the future and automatically push a related message to the digital sign. For example, ice cream is generally expected to sell less well in winter, but a sudden heat wave might have shoppers clamoring for the freezer aisle. The system would automatically identify sales as higher than the AI expected given all factors at play, triggering an alert to the digital sign, which in turn loads a graphic that tells shoppers that ice cream is selling particularly well today. Shoppers, recognizing the merit of their peers’ decision, follow suit. There is no need to spend time setting up or managing a certain promotion. The automated system instead succeeds just by reinforcing subconscious whims.

The important thing is that with a digital screen supported by AI, you don’t need to know why a trend exists. You just need to know it does exist.


One other unique content-adaptive feature of AI is the ability to deliver far more impactful, dynamic and personalized newsletters. First, AI allows customers to receive newsletters with content that meets their expectations and contains content they’d prefer to engage with. But the really interesting effect comes once customers start opening the content.

When people start clicking, the content hits the AI engine, which can analyze and update their responses to the content in real-time based on what they click on and spend time reading. Then, for future readers, the AI can instantly remove or reduce the prevalence of underperforming content, raise the profile of high-performing messages and improve both customer experience and ROI.

Automated, Real-Time Relevance

While those who ignore history are doomed to repeat its failures, only those who can make intelligent predictions about the future can adjust optimally to real-time environments on the fly. And that’s the promise of the GK Software AI-enabled platform. With our AI engine, real-time learning based on behavior looks like a chess game with a decision tree. The platform learns the moves ahead and uses predictive analytics to anticipate future product and consumer behavior.

To act based on historical data is a stable business approach. But to use our AI to predict outcomes and adjust to discrepancies in real time is the true power of AI in the store.