AI in retail: automation, forecasting and planning
Artificial intelligence (AI) has long ceased to be a futuristic concept and is now actively used in retail. It not only automates routine processes but also enables more precise decision-making based on data analysis. Thanks to AI, retailers can forecast demand, optimise stock levels, adjust product placement, and even personalise offers for customers.

However, it is important to understand that AI is not a magic wand that solves all problems with the push of a button. Its effectiveness depends on the quality of data, the right algorithms, and proper implementation. If a retailer neglects data analysis and system configuration, the technology will not deliver the expected results.

Where is AI already actively used in retail?
  • Demand forecasting considering promotions, holidays, and external factors
  • Automatic stock replenishment without employee involvement
  • Optimisation of product placement based on customer behaviour
  • Automatic product substitution and adjustment of the assortment matrix
  • Analysis of promotional campaign effectiveness and personalised offers

AI has already proven its effectiveness in both large retail chains and small shops. The question is how well businesses are prepared to harness its potential.

How artificial intelligence is changing retail

AI in retail is not just a trend but a real tool for improving business efficiency. Its primary function is to analyse vast amounts of data and use the insights to optimise key processes, from ordering and logistics to product placement and personalised marketing.

In the past, stores operated under the principle of "the more products on the shelf, the better." Now, AI enables a more targeted approach:
  • Optimising stock levels to prevent shortages or overstocking.
  • Placing products in a way that maximises profitability.
  • Automating routine processes to reduce the workload on staff.
Speed and accuracy of forecasts
AI analyses not only sales data but also external factors such as weather, holidays, trends, and competitor activity. This allows for more accurate predictions compared to traditional sales-based forecasting.

Personalised approach to each store
Every store operates at its own pace. In one area, fresh vegetables and fruit may be in high demand, while in another, ready-made meals are more popular. AI helps adapt the product range and shelf layout to match real customer needs.

Automation of ordering and product replacement
AI eliminates the need for employees to manually create orders. It takes into account demand, stock levels, and seasonality while also suggesting replacements for temporarily unavailable products.

Flexibility in assortment management
AI enables real-time assortment adjustments: if a product is selling worse than expected, the system suggests replacing it with a more in-demand alternative.

What exactly does AI change?

Time savings – less routine work for staff.

Reduced losses – accurate forecasts minimise write-offs and stock shortages.

Sales growth – optimised product placement and personalised offers increase the average transaction value.
Why should businesses implement AI?
AI is already transforming retail, and companies that fail to leverage its capabilities risk losing their competitive advantage.

Demand forecasting: accuracy and data

Traditional demand forecasting methods rely on analysing past sales. However, this is not enough. Customer behaviour is constantly changing, influenced by holidays, promotions, weather, trends, and even events like World Cup tournaments. Artificial intelligence takes all these factors into account, providing more accurate forecasting and more efficient inventory management.
AI takes into account not only internal data (sales, stock levels, seasons) but also external factors:

  • Holidays and weekends (e.g., increased demand for sweets during New Year celebrations).
  • Promotions (how did past discounts affect sales?).
  • Weather (e.g., higher demand for ice cream in hot weather).
  • Social trends (influencer impact, the popularity of healthy lifestyle products, etc.).

Unlike traditional models, AI operates in real-time, continuously analysing changes. If the forecast for a particular product deviates from the norm, the system automatically adjusts orders.

How does AI-based demand forecasting work?

Order accuracy – fewer stock shortages or surpluses.
Flexibility in assortment management – quick response to demand changes.
Cost savings – reduced losses from unsold products.

AI not only forecasts but also adapts to demand changes, helping retailers stay one step ahead.

How does this benefit retailers?

Automatic ordering: eliminating the human factor

Order processing is one of the most critical and labour-intensive tasks in retail. A miscalculation can lead to either stock shortages (resulting in lost sales) or excessive purchases (tying up working capital). Artificial intelligence solves this problem by automatically analysing demand and generating highly accurate orders.
Challenges of manual order processing
How AI improves the ordering process
  • Human error – a manager may miscalculate order quantities
  • Labour-intensive – data analysis, form completion, and approvals take a lot of time
  • Lack of flexibility – manual ordering does not always account for seasonality and sudden demand changes
  • Analyses sales and stock levels in real-time
  • Considers promotions, holidays, weather factors, and trends
  • Adjusts orders based on demand fluctuations
  • Automatically prevents warehouse overstocking

Why is automatic ordering better than manual ordering?

  1. Data collection – the system analyses stock levels, sales rates, and external factors.
  2. Order generation – based on forecasts, AI calculates the optimal purchase volume.
  3. Adjustment – if demand changes suddenly (e.g., due to a promotion), the system adapts the order.
  4. Supplier submission – after verification, the order is automatically placed.

How does automatic ordering work in practice?

Automatic product substitution: what to sell when the required item is unavailable

Every retailer faces situations where a certain product is temporarily unavailable. If a customer cannot find their usual product, they may choose to shop elsewhere. Artificial intelligence solves this problem by automatically selecting alternative products and adjusting shelf placements.
AI analyses the assortment and suggests alternatives based on the following criteria:

  • Product category (e.g., a similar type of toothpaste).
  • Brand (if one brand is unavailable, a similar one in the same price range is suggested).
  • Price (products within the same price bracket are recommended).
  • Popularity (AI considers which alternatives are most frequently purchased).
  • Consumer preferences (e.g., if milk chocolate is out of stock, the system may suggest dark chocolate).

The system not only identifies suitable substitutes but also adapts the planogram to help customers quickly find an alternative product.

How does AI select product substitutions?

  1. Tracking unavailable items – the system identifies which products are temporarily out of stock.
  2. Finding the best replacements – AI analyses alternatives based on price, category, and customer preferences.
  3. Updating shelf layouts – the system adjusts the planogram so that customers see available products first.
  4. Sales analysis – AI monitors the effectiveness of substitutions to optimise future recommendations.

AI-driven product substitution

Automatic product substitution is not just about convenience; it is a way to retain customers. If shoppers can always find a suitable alternative, they are less likely to go to a competitor, which helps maintain conversion rates and increase customer loyalty.

Why is this important?

Personalised planograms: how AI adapts product layouts for each store

Traditional planograms are created manually and rarely consider the unique characteristics of a specific store. However, customer behaviour can vary significantly even between two locations within the same retail chain. Artificial intelligence solves this issue by generating personalised planograms tailored to store layout, demand, and customer preferences.
AI analyses:
  • Sales data for the specific store – identifying which products are popular and which are overlooked.
  • Available space – considering shelf sizes, racks, and aisle layouts.
  • Seasonality and promotions – adjusting displays based on current trends.
  • Customer behaviour – tracking how shoppers move through the store and which products catch their attention.

Using these insights, AI automatically generates planograms that are not just visually appealing but truly effective in driving sales.

How does AI optimise product placement?

Challenges of traditional planograms
  • Uniform product placement across all stores, ignoring local differences
  • Time-consuming and requires manual adjustments
  • Does not adapt to changes in demand
Advantages of AI-driven planograms
  • Automatically consider demand, store layout, and customer preferences
  • Quickly adapt to changes (e.g., introduction of new products)
  • Increase sales by optimising product placement for convenience

Why are personalised planograms better than standard ones?

  1. Data processing – analysing sales, store layout, and customer movement patterns.
  2. Generating the optimal layout – AI creates a planogram considering all key factors.
  3. Implementation and monitoring – product placement is adjusted in real time.
  4. Performance evaluation – results are analysed, and the system makes further improvements.

How does it work in practice?

Challenges of AI implementation: what to consider before launch

AI can significantly enhance retail processes, but its successful implementation requires time, accurate data, and proper configuration. Mistakes at this stage can reduce effectiveness or even lead to negative outcomes.
Inaccurate or incomplete data
AI relies on existing system data. If sales records, stock levels, or customer behaviour data are disorganised or contain errors, the algorithms will produce inaccurate forecasts and recommendations.
❐ Solution: Conduct a data audit, eliminate duplicates and errors, and establish unified data standards.

Employee resistance
Store staff and managers may fear that AI will replace their jobs or simply not trust the algorithms. Without team support, the implementation of new technologies can be delayed.
❐ Solution: Train employees and explain that AI is a tool to assist them, not replace them.

Unoptimised business processes
If a company already has issues with logistics, inventory management, or data exchange between departments, AI will not fix them—it will only highlight them.
❐ Solution: Review and address existing inefficiencies before implementing automation.

Lengthy adaptation period
AI does not work perfectly from day one. The first few months involve testing, fine-tuning algorithms, and collecting feedback.
❐ Solution: Initially, use AI in a recommendation mode, allowing human oversight for final decisions.

High expectations
Some companies expect an immediate increase in sales and cost savings. However, AI is not instant magic; it is a tool that improves over time.
❐ Solution: Implement AI gradually, analysing results at each stage and adjusting the strategy accordingly.

What challenges can arise when implementing AI?

  • Start with a pilot project – test AI in a few stores before full-scale implementation.
  • Prepare the data – without high-quality information, AI will not function correctly.
  • Train employees – engage the team and explain how to use new technologies.
  • Evaluate effectiveness – monitor AI results and adjust settings as needed.

AI is a powerful tool, but only when implemented correctly. Companies that take a thoughtful approach achieve real results: increased sales, reduced costs, and an improved customer experience.

How to minimise risks?

AI has already proven its effectiveness in the retail industry. Companies that implement it today gain a competitive edge by improving forecast accuracy, automating orders, and creating personalised planograms.

Where to start with AI implementation?
✔ Assess current processes – identify which tasks can be automated.
✔ Prepare your data – ensure accuracy and completeness.
✔ Launch a pilot project – test AI in selected stores before scaling.
✔ Train staff – educate employees on how to work with the new system.
✔ Analyse results and refine strategy – AI becomes more effective with continuous optimisation.

AI in retail is no longer an experiment but a standard. Companies that ignore it risk losing their market position, while those that implement it strategically benefit from increased sales, lower costs, and higher customer loyalty.

Conclusion: why AI is not the future but the present of retail

Tilda Publishing