Breaking News

Why AI and ML are Key Ingredients for Quick-Service Restaurants

Despite the surge in takeout dining since the start of the pandemic, the quick-service restaurant industry is struggling. Labor shortages, supply chain problems, reduced in-store traffic, and rising costs have pushed the industry into a crisis. Coping with these headwinds requires a new approach to controlling costs, managing available labor, and improving customer experience. Artificial intelligence (AI) and machine learning (ML) technologies can help by leveraging data to optimize quick-service processes and operations. Below are some of the key use cases where these technologies can help the industry. 

Improving the drive-thru experience with AI/ML 

Most fast-food customers have had the frustrating experience of getting the wrong bag at the drive-thru window. In September 2021, only 85 percent of quick-service drive-thru orders were accurate, down from 87 percent in 2020, but an AI/ML system can reduce drive-thru errors by as much as 90 percent.  

READ MORE: The drive-thru of the future arrives ahead of schedule

AI/ML technologies can be trained to learn menus and can use natural language processing (NLP) to recognize speech and correlate voice orders. A well-trained system will lead to fewer misunderstandings and data entry errors. A more advanced AI/ML system can recognize customers’ faces to generate personalized order prompts for faster ordering, a better customer experience leading to more upsells. At least two major chains are already working on such systems, which will have to learn a range of accents and languages to be effective.  

Matching quick-service restaurant inventory to demand

AI/ML technology can also optimize food purchasing, delivery, and inventory management to reduce waste, manage costs, and avoid shortfalls during peak demand times. This is critical in an industry where food costs have risen for most operators, and where 96 percent of companies experienced supply delays or shortages in 2021.  

For example, the system can monitor ingredient and supply inventory levels by drawing on point-of-sale (POS) system data in real-time, learning to flag items for reorder based on sales. AI/ML solutions can also learn to compare sales to ingredient inventory to ensure that employees are using the correct portions for recipes, resulting in better quality meals and less waste. Over the longer term, machine learning can accurately predict demand increases related to holidays, major events, and even changes in the weather, so stores will have the right amount of ingredients and supplies on hand. 

Optimizing employee scheduling with machine learning

Staffing is a major expense for restaurants, and many brands are struggling to hire and retain employees. According to the National Restaurant Association’s 2022 State of the Restaurant Industry Report, 78 percent of operators lack the staff to meet demand, and there are more than 1 million unfilled positions in the industry.  

AI/ML technologies can be trained using historical labor information across different stores by considering a multitude of factors like skillset of labor, peak demand hours during the day and months of a year, special promotions and so on. A well-trained AI/ML application can benefit Quick-serves by helping them predict labor needs based on demand to optimize employee hours and labor spend accordingly. Predictive scheduling can also benefit employees by ensuring they’re on the schedule when they’re needed and giving them ample notice of their work hours. This straightforward scheduling can help improve employee satisfaction and retention.  

Gathering better insights for pricing, promotions and discounts

As brands work to manage costs amid inflation, machine learning technology can help them implement dynamic pricing strategies based on fluctuating costs for ingredients, nearby competitors’ prices, and predicted customer price tolerance.  

That said, dynamic price adjustments don’t always need to be upward. An AI/ML system that’s tracking inventory levels and stocking dates can also suggest discounts or promotions on menu items when there are overstocks of perishable ingredients, to prevent waste. This can be especially beneficial given that 931 million tons of food is wasted at retail and consumption stages and the average cost associated with food waste is around 5.6 percent of total sales. 

Maintaining kitchen equipment more efficiently 

When kitchen equipment breaks, the restaurant’s top and bottom lines take a hit. Broken specialty equipment like espresso and soft serve ice cream machines can require long wait times for repairs and generate negative word-of-mouth. This is where Internet of Things (IoT)-enabled appliances can help. These appliances can send key data to a centralized data cloud. AI/ML technologies can then leverage this data to understand past breakdown parameters and anticipate future breakdowns, enabling quick-serves to conduct predictive maintenance rather than preventive maintenance or repairing the appliance after it is broken. 

Building an AI/ML system for quick-service restaurants 

Developing the kind of smart systems that can deliver these benefits requires a top-down commitment at the corporate level, a champion who will sell the AI/ML concept to internal stakeholders and identify an implementation strategy. Starting small, with pilot programs that can deliver easy wins from a cost-benefit perspective, can ensure success and generate more support and buy-in.  

With pilot programs identified, companies will also have to consider staffing requirements. For example, a typical quick-serve may not have data engineers and scientists, Python developers, or AI/ML experts in house, so they’ll need to source that talent. Once the right team is in place, it’s time to start unifying and standardizing data from POS systems across locations, apps, and online ordering systems. It’s also important to consider where the data will be stored and how it will be accessed.  

When the data is ready, it’s time to build models to train machine learning algorithms. This requires time, testing, and refinements to reach an ideal accuracy range of 90 percent or better. At that point, the system is ready for small-scale testing on real data before expanding to a pilot program, where the time, testing, and refinement process repeats until the pilot is ready for a wider rollout and perhaps more use cases. 

The big takeaway for AI and ML applications in the quick-service restaurant space is that we’re only at the initial stage of a development that can offer immense benefits, which is why major players are putting time, money, and effort into building these systems. As more companies deploy AI/ML systems, their competitive advantages will accelerate, so quick-service leaders that want to keep up in the crowded market should start planning their own AI and machine learning programs now—or risk falling behind in terms of cost savings, sales growth, employee experience, and customer experience.  

Karthik Suryanarayanan is Principal of Digital Customer Experience at Capgemini Americas. He has over 18 years of digital consulting experience across sales, service, marketing, and commerce within high-growth organizations covering a unique mix of technology expertise, digital strategy, creative design, and digital technology enablement in both domestic and international markets. In addition, he also helps clients in leveraging new trends in the industry like Internet of Things (IOT), Artificial Intelligence(AI), Machine Learning(ML) and Robotic Process Automation  (RPA) 

Shibu Abraham is a Managing Applications Consultant at Capgemini Americas. Shibu has been in the consulting industry for 11 years and the IT industry for 22. He has been focused in the QSR space for the last five years working with some of the largest brands in the industry. Shibu is based in Chicago, Illinois.