Once the exclusive territory of science fiction, artificial intelligence is starting to penetrate many retailers’ go-to-market strategies. And for good reason.
The technology stands to radically change the face of retail. By combining huge data sets and massive amounts of processing power, advanced machine learning algorithms can automate time-intensive functions while reducing human error and improving the customer experience.
However, implementation represents a formidable task that requires long-term thinking, strategic planning and technical expertise. While adopting AI may seem daunting, grocers can make it easier by first hiring the right people, organizing their data to support the technology, and carefully evaluating vendors so that they can realize the power of intelligent decision automation.
Empowering the Right People
Just as corporate executives are beginning to see the vast potential of AI to transform daily processes and decision-making, they are also realizing they simply do not have enough qualified talent essential to driving the enterprise-level changes they seek.
Since December 2013, job postings for data science roles have skyrocketed by 256%. And though data science is quickly gaining popularity with students seeking to enter the field, there remains a series backlog of talent. The talent gap is especially acute among those with significant retail or grocery experience, who are in even shorter supply.
This shortage of data scientist talent means many business leaders are assuming analytics roles, including senior executives who are using their practical expertise to guide the implementation of AI across the enterprise. These leaders bring a unique skill set to the table with their years of industry experience and insider knowledge of what processes can benefit the most from intelligent decision automation.
However, without the right technical skills, these leaders can typically only make a finite amount of progress without seeking additional help. This makes the selection of an AI vendor with retail experience among the first steps grocers need to take after they have empowered transformation leaders.
Thankfully, the rising tide tends to lift all ships. How does a data analyst within retail grow to become a data scientist that can be functional within a retail environment? Working with vendors or consultantsthat intimately know the space can help internal employees become more capable.
In addition, outside experts will often be more familiar with the latest technology, can offer technical recommendations to speed along AI development and maximize and empower existing talent to tackle the huge challenge of formatting data for AI use.
Partner With the New Stack
The future of retail is in the cloud. Not just in terms of consumer apps that enable new fulfillment strategies but cloud providers that offer cutting-edge, at-scale technologies and startups that can co-develop solutions to match customer needs.
Traditional on-premise offerings from companies such as IBM, SAP and Oracle can only take grocers so far. In contrast, cloud infrastructure providers offer an elastic foundation for grocers to continuously improve their processes as their needs grow.
Specifically, grocers should seek out partnerships with cloud providers such as Microsoft and Google because these are the companies—working side by side with startups—that are building the next-generation of machine learning platforms and new vertical AI applications. Only by investing in the cloud will grocers be able to compete with the largest players, such as Amazon, Kroger and Walmart, that are already using AI in innovative ways, such as inventory management and cashier-less commerce.
Investments in machine learning and AI require scalable solutions, and when building their new solution stacks, grocers should look to the cloud to manage their immense data and processing requirements.
Organizing and Cleaning Data Is Hard
Many grocers do not initially realize just how difficult it can be to set up working AI models. It often requires thousands of data sets that first need to be cleaned and formatted before they can be used. Furthermore, data cleansing is among the most tedious, time-consuming tasks for data scientists, and even after it’s cleaned, it still needs to be formatted and organized.
That task is magnified by the grocery industry’s continued reliance on legacy systems. ERP or point-of-sale systems, for example, have a tendency of locking away data behind hard-to-use interfaces that can’t easily communicate with newer systems. It can be difficult to extract data from these sources and further adds to the time it takes to set up working models that can automatically generate intelligent decisions.
So after grocers have empowered employees to test AI deployment, the next step is developing a clean layer of data that consolidates all of these various sources into a unified platform that can be used for analysis and, ultimately, automation.
Unlocking the Power of AI
The effective selling price of an item when it’s sold is driven by a calculation of actual cost, promotion cost, vendor subsidy, regional discount, plus a number of other factors. For that reason, it’s not enough to look simply at POS data to create AI-driven recommendations. Grocers must look at the bigger picture to transform their operations and unlock the power of intelligent decision automation.
Doing so requires stitching together dozens of data sets and millions of unique data points in a way that makes sense. Once grocers have connected data from their legacy and cloud systems into a comprehensive view that goes beyond sales and price, the next step is applying advanced machine learning techniques to identify patterns.
Reviewing months or years of data across connected systems, grocers can finally apply intelligent decision automation to solve the unique business challenges identified by their internal leaders. For example, using AI analysis, grocers can run advanced models to test how changes in pricing and promotion in certain markets would impact sales and make intelligent decisions to optimize lift. Similarly, they can determine how much to ask for in vendor subsidies, or find improvements in the accuracy of their demand forecasting to reduce their number of stockouts.
Smart AI applications that can deliver real business results like this might seem far away, but the truth is that they are already being deployed by the biggest operators. AI isn’t science fiction anymore, and by hiring the right people and optimizing their systems, it’s time the rest of the grocery industry caught up.
Kerry Liu is the co-founder and CEO of Rubikloud, an AI and machine learning platform for enterprise retailers.