Call Today: 832 338-2926
Call Today: 832 338-2926
This case study examines how generative Artificial Intelligence (AI) transformed the relationship between Amman Express, a Quick Service Restaurant (QSR), and its franchisees. Traditional data analysis methods were slow and inaccessible for Amman Express franchisees, hindering informed decision-making. The data analytics team at Amman Express was overwhelmed with requests. This essay will explore how implementing a generative AI solution using a Large Language Model (LLM) enhanced data accessibility, streamlined operations, and improved overall business efficiency and strategic decision-making.
Quick Service Restaurants (QSRs) are a mainstay of modern life, offering convenient, consistent, and budget-friendly meals to individuals and families. Examples include McDonald's, Starbucks, and Pizza Hut. QSRs prioritize offering customers a familiar dining experience with consistent menus, service, food quality, and taste regardless of location. The global market size for QSRs was valued at approximately $2.1 trillion in 2021, with revenues in the United States estimated to be $382 billion in 2022. The industry contributes to local economies through job creation, wages, and spending on supplies and services.
The QSR industry is highly competitive, with established brands vying for market share through innovative offerings, competitive pricing, and superior customer service. Brands must quickly adapt their menus, prices, and service models to keep ahead of evolving customer expectations and preferences. During the Covid-19 pandemic, successful QSR businesses quickly ramped up kiosk, phone app, and online ordering models, adapted their menus and packaging for delivery, and offered a wider range of family meal deals. QSR companies also face increasing competition from non-traditional competitors, such as convenience stores and grocery stores offering prepared meals and meal delivery services.
The franchise business model is a mainstay of the QSR industry. Used by the majority of QSR brands, it allows individuals (franchisees) to operate a restaurant under an established brand (franchisor) like McDonald's or Subway. Franchisees pay an initial fee and ongoing royalties for the right to use the brand name, standardized operating procedures, and established supply chains. In return, franchisors offer training, marketing support, and ongoing guidance. This model allows rapid brand expansion while providing franchisees with the support and brand recognition of a successful business. Key to a successful relationship between a QSR franchisee and corporate headquarters is communication and collaboration.
Over the past five years, technology has impacted the QSR industry significantly. Mobile technology has enabled contactless ordering, payment, and delivery, as well as hyper-customized menu options, location-free brands, and innovative loyalty programs. Technology solutions are also being explored to address the post-Covid issue of finding and retaining onsite staff, ranging from self-ordering kiosks to robotic fry cooks. Data analytics, cloud computing, and artificial intelligence are revolutionizing how QSRs can transform their vast data repositories on customer behaviors, supply chains, and operations into actionable, lucrative innovations.
Communication and collaboration between a brand and its franchisees are crucial to the success of both parties, but rarely easy or straightforward, particularly in the sharing and analysis of data. This case study focuses on how applying generative AI affected the relationship between Amman Express, a QSR brand, and its franchisees, and what the resulting impact was on the overall business.
Amman Express is a QSR brand that focuses on providing fresh and healthy meals with a Mediterranean twist. Their products include salads and bowls, flatbreads, and traditional Middle Eastern desserts. Founded in Dearborn, Michigan, the brand has grown its presence in the United States to over 2,000 restaurants and has expanded into Canada. Franchisees owned 79 percent of Amman Express's locations at the end of 2023. The company's stores are located primarily in urban and suburban areas, with customers using onsite kiosks, the brand's smartphone app, or home delivery provider apps to order.
Amman Express gathered an overwhelming amount of data relating to their business, including revenue, supply chain, customers, rewards programs, operations, geographical markets, and competitors. However, this data was housed in various repositories developed by different vendors, making it difficult to bring together for analysis. The company invested in a small data analysis team responsible for locating, analyzing, and creating reports from the company's many databases. The team was also responsible for cleaning, organizing, and normalizing data in disparate databases.
For Amman Express franchisees, requesting data and reports relevant to their businesses involved several steps and often resulted in frustration. Franchisees could request a small selection of "canned reports" or submit a query via email to the data analytics team. However, the queue for reports was long, with an average time from request to response of 4-6 months. Franchisees complained that by the time they received their reports, the need or opportunity had passed, resulting in lost revenue, supply chain cost overruns, or lost market share.
Corporate strategists, product owners, and stakeholders also experienced frustrations due to the size of the data analytics team and the balkanization of data repositories. Recent, detailed information from franchisees regarding product performance, marketing campaign feedback, or supply chain issues was hard to come by. Menu developers and regional marketing directors relied on incomplete data or word of mouth, potentially leading to missed opportunities.
With the installation of a new CEO in 2021, Amman Express entered a new phase focused on being first to market and increasing market share. The new CEO advocated for an overhaul of the company's approach to technology, installing a new CTO and promoting the early adoption of cutting-edge technologies to solve previously recalcitrant problems.
A cross-functional task force conducted a root cause analysis and concluded that the data analytics team was not sufficiently resourced to handle the number of queries, the data itself lacked significant structure and consistency, and the report generation process was executed almost exclusively by hand. To make the data reporting system work, it would have to be automated.
Amman Express leadership chose Axiom Solutions, an outside IT consultant, to identify and implement an automated solution. Axiom proposed a solution built on generative AI, using a Large Language Model (LLM) specifically engineered to understand massive amounts of text data and perform language-based tasks.
Axiom proposed that Amman Express stand up an LLM behind the company's firewall and give it context of their internal data. The LLM would form the core of a platform that franchise owners could ask questions of and receive accurate, easy-to-understand answers within seconds. Franchise owners could better understand their businesses, compare themselves against peers and regions, and better manage their supply chains using actual data.
The existing Amman Express data had to be cleansed and brought together in a data lake, a flexible, centralized platform for storing all types of data. Axiom worked with over eight Amman Express teams to migrate their data and add meta-tags, improving its usability.
The AI installation project on the Google Vertex AI platform involved containerizing the AI, ensuring its operability within the secured bounds of their Google Cloud instance. The subsequent phase involved developing an orchestration layer, empowering the AI to emulate a data analyst adept at addressing franchisees' data queries. The project culminated in an AI system that responded to franchisee inquiries with high precision and speed.
A training program for franchisees and corporate staff was rolled out 30 days before launch, focusing on introducing the new application, core AI and NLP concepts, interacting with the AI, customizing the dashboard, and providing feedback to the AI. Hands-on workshops and ongoing support ensured users were equipped to effectively use the application.
Impact on Franchisees: Franchisees gained significantly faster access to data, enabling them to make better business decisions. Initial skepticism transformed into enthusiasm as they engaged with data more confidently.
Impact on Data Analyst Team: The data analyst team was freed from routine query generation and data retrieval, allowing them to focus on deep-dive analysis and strategic initiatives. Analysts moved into a more advisory role, fine-tuning insights and guiding franchisees in understanding complex data trends.
Impact on Business Intelligence Team: The BI team became more agile, allowing them to collaborate more effectively with franchisees and tailor intelligence reports to address specific needs.
Impact on Overall Business: Streamlined decision-making, based on accurate and swift data interpretation, became the norm in several key areas. Franchisees, with direct access to customized insights, could drive local innovations. A follow-up internal study found tighter alignment between operational efficiency and strategic goals, with the AI assistant playing a central role.
The successful implementation of generative AI at Amman Express highlights the potential of AI in the QSR industry. Key takeaways include empowering franchisees with data access and informed decision-making, enhancing data analytics by automating tasks and freeing up human expertise, improving collaboration between franchisors and franchisees, and creating dynamic business intelligence for faster, more responsive decision-making. As AI technology continues to evolve, QSRs that embrace its potential will be well-positioned to thrive in a competitive landscape.
Copyright © 2024. All rights reserved.