Chipotle’s recipe for digital transformation: Cloud plus AI

When Curt Garner became Chipotle’s first CIO in 2015, the only technology used for online restaurant ordering was, “believe it or not,” a fax machine, he says.

Seven years later, the Newport Beach, Calif.-based company is piloting a system with a robotic arm dubbed “Chippy” that prepares the fast-casual chain’s famed fried salt-and-lime chips homemade every day without human labor, with the aim of tailoring output to the daily needs of each restaurant.

“We are using robotic technology to do all the manipulation that a human would be doing in that process,” says Garner, noting that “Chippy” is currently in use in only one of Chipotle Mexican Grill’s 3,000 stores but use will expand. “We’ve spent several months in our lab refining the recipes and the processes to make it possible.”

The former Starbucks CIO didn’t waste time jumpstarting the restaurant chain’s digital transformation in 2016, applying and evolving his “learnings” from Starbucks to pilot a digital ordering system that would enable Chipotle customers to order online in advance and pick up their food from a drive-through window — without use of speaker boxes or menu boards typical of drive-through fast-food restaurants. 

Removing the physical speaker box on site was a simple concept but a key part of a bigger digital transformation Chipotle kicked off in 2018 that led to an explosion in business, in large part because the digital ordering system required less human labor during the pandemic. Chipotle’s digital business in 2022 was $3.5 billion — magnitudes bigger than the $100 million to $200 million annual average before digital ordering processes were implemented in the business.

Chipotle IT’s secret sauce

Garner credits Chipotle’s wholly owned business model for enabling him to deploy advanced technologies such as the cloud, analytics, data lake, and AI uniformly to all restaurants because they are all based on the same digital backbone. Many restaurant chains, in contrast, are individually owned and operated as franchises.

“We’re not 3,100 different companies but one restaurant 3,100 times, while each of those locations are impacted differently by their trade area, commute patterns, and weather patterns,” says Garner, who was promoted to CTO after proving his mettle. “Part of the strategy and policy that we put in place in early 2016 around the cloud was that the closer we were to the consumer experience, the more we could abstract from that consumer experience and leverage cloud and commodity services.”

And the cloud is at the center of all that. Chipotle’s digital commerce platform is Microsoft Azure, and its internal business processes such as ERP have been migrating from on-premises Oracle to Oracle Cloud. Because the company got an early start, 98% of Chipotle’s workloads now run on the cloud, Garner says, enabling the chain to retire its Denver-based data center and begin implementing a series of-next generation cloud-native applications to grow its business across the US and Europe.

That foundation has set up Chipotle well to not only survive but thrive through the pandemic, says Sandeep Unni, senior director analyst at Gartner.

“In an industry that has been historically resistant to change and operates on razor-thin margins, the importance of investments in digital, data and customer experience cannot be overstated,” Unni says. “I view Chipotle as a great example where strategic bets in digital transformation made well before the pandemic have paid off.”

Currently, Chipotle is exploiting a variety of cloud services that are part of the Microsoft Azure platform, such as its AI and ML modeling services. The restaurant’s Microsoft consumer data platform for customer ordering online makes use of embedded AI that performs tasks with no human intervention, but Garner is also heavily exploiting another AI application known as H20, which Chipotle has been using for two years.

“We really like their cloud-based AI program,” Garner says. “It allows us to build models and applications very quickly and with transparency. They’ve got a management plane on top of that, which allows us to dynamically change and monitor performance across the platform and its powerful set of tools.”

Building on the foundation

Garner also recently implemented RFID in 100 restaurants to help track its food supply from local farms to distribution centers to each restaurant. The system ties into Chipotle’s Oracle system, offering traceability, points of origin, and real-time insights into supply location, especially if suppliers have issues. But Garner sees further potential for the system.

“I can see a future where we combine technology called ‘real food print’ that we have implemented that allows the customer to understand the environmental impact of that order, such as the local farm their food came from,” Garner says, noting that the company sources more than 30 million pounds of locally sourced produce per year, something that is becoming more important to customers, especially college students.

Garner’s flexible cloud infrastructure, which includes a data lake built on Snowflake, has also facilitated the chain’s implementation of PreciTaste, a SaaS and AI system that enables each store to properly forecast its supply needs. The system provides “demand-based cooking and ingredient preparation forecasts to optimize throughput and freshness while minimizing food waste,” according to the company. “Leveraging AI and machine learning, the system monitors ingredient levels in real time and notifies the crew how much to prep, cook and when to start cooking, while automatically populating real-time production planning for each restaurant,” the company recently announced.

In spite of Chipotle’s heavy reliance on technology, Garner maintains that fully automating the stores with robots would hurt the business, as customers still want interaction with and understanding from human servers.

While this longtime CIO and his team are keen on using the most modern tools of the trade to serve the freshest food possible, Garner also makes it clear that he is not married to any one cloud or technology infrastructure.

“Technology is changing all the time, and one of the things that I appreciate about the cloud as a growth company, there’s not only the flexibility we have to rapidly scale up or scale down and provision more, but we’ve got the ability to move to other platforms over time without the cost of legacy that’s typically involved in on-premise solutions,” the CTO says.

“I never think of our architecture of static,” Garner adds. “I think flexibility and nimbleness are going to continue to be a competitive advantage for us for some time.”

Analytics, Artificial Intelligence, Cloud Computing, Digital Transformation