AI in retail? It won't be without work.
Every few years, a new disruptive technology comes along that promises to revolutionize more than just retail. Artificial intelligence is certainly not a technology that will magically solve our problems on its own, but its use can bring significant benefits, whether in the form of automated decision making or machine learning.
But it doesn't come without work! In fact, it's more true than ever that good quality input data is essential for AI applications.
Let's take a look at some examples of retail AI applications that have worked in the real world.
Event-based sales forecasting is much more accurate with AI
Quality sales forecasting helps improve space planning, automated ordering, and the supply chain, all of which can be greatly enhanced by AI. Every retailer is unique and has its own specifics that need to be taken into account to ensure that the predictions work properly and are accurate enough.
However, providing AI with data on inventory history and past and future events that affect sales is critical to future sales predictions.
Merely knowing sales history is not enough.
Factors such as past sales, promotions, holidays, or sports games have a significant impact on buying behavior. AI can evaluate this impact and provide much better demand forecasts on a daily basis for thousands of products across hundreds of stores.
Automatic ordering is one of the most useful uses of AI in retail
The traditional replenishment model, which relies on manual orders placed by store associates, has many drawbacks. It takes a lot of time, care, analytical skills and experience to create a quality order.
With good demand forecasting and integration with planograms, AI can automatically create orders of higher quality than the average store manager could manage. The benefits of implementing automated or at least semi-automatic ordering include huge time savings for staff, improved product availability and customer satisfaction, and increased sales.
However, the introduction of automatic ordering is not without its challenges. Especially in the first few months, a lot of time and effort needs to be spent on testing and overall setup, so that the AI has a good idea of what we actually want from the days and reasonable limits to prevent unwanted anomalies.
Automatic replacements help update planograms and predict news
Another interesting application of AI is automatic product replacement. AI makes it easy to determine the right time and conditions to replace product A with product B. Simply define similarity rules, such as identical brand, closest price point, similar color, etc., and AI will determine the best available substitute and make the display change in each store individually.
Defining the similarity rules is critical to how the whole process works and can take weeks to months of human work, depending on the complexity and diversity of the assortment.
The applications are many – from automatically replacing unavailable products in planograms to predicting sales for a new product based on the most similar products that have sold in the past. All fully automated, on a daily basis, for thousands of products in hundreds of stores.
Thanks to AI, we can create Store Specific Planograms
Artificial intelligence has brought a breakthrough approach to the traditional planogram creation process. It makes it possible to generate thousands of optimized, store-specific planograms that are tailored to the available space in stores based on demand forecasts, store priorities, promotions, and many other parameters. It is no longer a problem to easily and quickly create planograms tailored to local specifics and have control over the entire implementation process.
Planograms can now be created using visual rules that reflect the customer's decision tree. Artificial intelligence takes care of the rest, creating planograms according to the specified visual logic so that more important products get priority. This is all based on sales and the available space and assortment in a given store. Even two stores with the same space can have slightly different planograms if customer behavior is different.
There is no doubt that the future of planogramming will continue to lie in the use of artificial intelligence combined with detailed data.
Why use AI-enabled technologies?
The concepts outlined above are proven real-world practices for many retailers and can significantly reduce overstock, increase sales, save store staff time and, most importantly, improve the customer experience. The use of AI in retail is undeniable – it opens the door to achieving results that would not otherwise be possible.
Written by
Petr Kavánek
CEO / Co-owner | Quant Retail s.r.o.