Dec 6, 2019

Collaboration success stories from within the community

By Patrick Quinn

Tuesday afternoon of Innovation Outreach: Boston began with a series of collaboration success stories from innovators within the community. Amy Thompson discussed Nestlé’s Digital Immersion Program, which has helped foster a culture of entrepreneurship at the multinational company with more than 300,000 employees. Later, David Frazee of Corporate Research Systems Lab (CRSL) shared a notes and highlights from a recent collaborative hackathon. 

But among the many remarkable success stories that were shared both in public and private throughout the three days in Boston, the collaborative prototype experiment that Ryan Bruels and Andy Lee from Starbucks shared certainly stood out. 

Why Partner Success Means Customer Success at Starbucks 

Bruels and Lee began their presentation by giving some context on the company culture at Starbucks. Bruels, who is the Director of Emerging Technology at Starbucks, and Lee, the company’s Innovation Product Manager, explained that the purpose of their division in the company is to “accelerate the creation of new compelling partner and customer experiences and business capabilities through radical collaboration, rapid experimentation, and emerging technology.’ 

Bruels proceeded to explain how the baristas at each Starbucks café are the backbone and lifeblood of the entire organization. This belief is affirmed by the fact that Starbucks refers to them as partners, and it is their job to engage with customers to prepare coffee drinks and the rest of Starbucks’ offerings, suited to each customer’s preference. 

The interactions between customers and partners are the most critical part of the entire business operation, Bruels explained. Therefore, the better positioned partners are to succeed and to be happy while on the job, the better each customer experience will be. Bruels touted the many ways that Starbucks already sets up partners to succeed, with a positive work environment and a broad portfolio of benefits that includes a college achievement plan, a stock program, and even transportation to and from work. 

So Bruels, Lee, and team sensed there were still ways to improve, particularly in the job itself. The Starbucks Emerging Technology team embarked on a mission: How can we set up our partners for even more on-the-job success?

Identifying the Biggest Pain Points and Conceiving a Solution 

Like true experts in design thinking, Bruels, Lee and team did not rely on their own judgments and beliefs in order to identify opportunities for improvement in the work experience of Starbucks partners. Instead, they leaned on the feedback and observations from the partners themselves. 

It became readily apparent that working the drive-thru window at Starbucks is among the most stressful parts about the job. It requires a great deal of multi-tasking: taking the customers’ orders, relaying the order to the rest of the staff, collecting payment and giving change. Above all, Bruels, Lee and team recognized that just the task of understanding customers’ orders proved a major challenge. 

And so the team came up with a novel prototype idea: What if they implemented a system that could interpret the orders of every customer and automatically relay it to the barista team, thus freeing up partners and saving them from an immensely arduous and frustrating task? 

Project RAMSAY: A Microsoft and Starbucks Voice AI Prototype Collaboration 

The Starbucks Emerging Technology team decided to move forward on this concept. They enlisted the help of the Microsoft's Commercial Software Engineering (CSE) team to develop the solution dubbed “Project Ramsay.” 

The goal was to create a hands-free and stress-free experience for partners at Starbucks cafés, specifically through in-store voice AI platform that could interpret a customer’s order and relay it directly to the barista team. A natural language processing (NLP) system like this would, in theory, free up partners to better coordinate and control the activity going on in their shops, to make for a much less stressful and much more positive work environment. 

Bruels laid out the three phases of Project RAMSAY: first, prototyping with baristas, followed by a functional and technology test, and finally, a data-collection phase from an in-store pilot. 

In the prototyping phase, the Starbucks and Microsoft collaboration team worked with barista partners to establish four critical requirements for the voice AI solution. First, the solution should be partner-facing, and should be responsive to the pace of each operation. Second, it should give partners complete control to override the NLP system and attend to customers in the best way they see fit. Third, it would have to give both visual and audio feedback so that partners can follow customers’ orders in real time. Finally, it would have to allow partners to maintain normal operations if and when an order was too difficult for the technology to accommodate. 

In the functional and technology testing phase, Bruels and his team mapped out the logistics of putting an NLP voice system to work. Developing an NLP system like RAMSAY was, of course, not without its challenges. For one, customers do not all order their cappuccinos, macchiatos, and pastries in a uniform way. This required the team to collect copious amounts of data to map out all of the possible ways for customers to express all of the nuances of their specific orders.

Once the multi-step system was tested out and deemed functional by the Emerging Technology team, it was ready for the final prototyping phase, the store data collection pilot. Bruels diagrammed how the cloud was leveraged in conjunction with the voice AI platform to enable a much more efficient order-collecting process. 

Rather than having a barista listen and interpret orders from each customer and enter them into a system to then later be prepared, leveraging the cloud with RAMSAY meant that partners in each Starbucks café could in real time correlate the audio collected with data on hand to efficiently interpret and process the order of each customer. One partner involved in the prototyping process expressed her enthusiasm for leveraging the RAMSAY solution. “Can I have this in my store?” she asked. “I will miss not having it.” 

Above all, Project RAMSAY is a lesson in what can be possible when collaborative innovation is approached in the right way. The Starbucks and Microsoft team's prototype succeeded not just because it reflected a commitment to make a major business improve, but rather because people were put at the center of the solution.  

By Patrick Quinn

Patrick Quinn is a member of the Innovation Outreach editorial team.