Key Takeaways

  • Starbucks plans to recognize participating drive-thru customers and present personalized digital menus.
  • The company will test AI-assisted drive-thru ordering in 2027 and is exploring AI ordering in its app.
  • Clear consent, restrained data use, and access to employees could determine whether customers view the experience as helpful or intrusive.

Starbucks ($SBUX) is preparing to bring more personalization into one of its busiest customer touchpoints: the drive-thru. The CEO stated the company plans to recognize participating customers as they arrive and show them personalized digital menus, part of a wider push to apply AI across ordering and service.

The wording matters. The CEO described "experiences that recognize participating customers as they arrive," indicating that the feature is intended for customers who choose to participate. Starbucks has not specified in the provided announcement how that recognition will work, what information will identify a customer, or how long related data will be retained. Those details could shape public reaction.

Starbucks will also begin testing AI at drive-thrus in 2027, where the technology is expected to help baristas take orders more accurately. Separately, the company is exploring AI-based ordering in its app. Together, those projects suggest a system that could connect customer preferences, order histories, app activity, and real-time ordering assistance. The practical goal is straightforward: fewer errors, faster decisions, and less friction at the speaker.

Personalization can provide value without feeling intrusive. A menu that foregrounds a customer's usual drink, remembers a preferred milk option, or avoids irrelevant promotions can save time. But what happens when a menu appears to know more than the customer expected? Recognition becomes uncomfortable when the data source is unclear, the recommendation feels manipulative, or participation is difficult to disable.

That tension extends well beyond coffee. Salesforce Einstein, Zendesk AI, and Intercom Fin reflect a broader move toward automated intent detection, tailored responses, and AI-assisted customer service. Forrester has likewise framed B2C marketing as undergoing a wider transformation as companies rethink how customer data, technology, and interactions fit together.

Trust, however, remains uneven. The IDC/SAS Data and AI Impact Report found that 62% of organizational respondents identified data privacy as a major generative AI concern. Transparency and explainability followed at 57%, while 56% cited ethical use. Although 78% said they trusted AI, only 40% had invested in governance, explainability, and ethical safeguards that could make that trust demonstrable.

For Starbucks, a credible rollout would make several things visible before customers reach the ordering screen. The experience should identify when AI is involved, explain what enables recognition, disclose which information shapes recommendations, and provide an accessible opt-out. Sensitive data can be minimized, while customers should retain a simple path to a barista when automation misunderstands an order. Convenience loses much of its value if correcting the system takes longer than ordering normally.

The NIST AI Risk Management Framework offers a useful structure through its Govern, Map, Measure, and Manage functions. Applied here, Starbucks could define accountability, map potential customer harms, measure accuracy and opt-out behavior, and manage problems as testing expands. Accuracy alone is not enough. Teams could also monitor whether recommendations pressure customers, whether recognition produces false matches, and whether certain users encounter more ordering errors.

Done carefully, Starbucks' approach could make drive-thru service feel more attentive without asking baristas to remember every returning customer. Done opaquely, it could turn an ordinary coffee stop into an unexpected demonstration of data collection. The dividing line is less about how sophisticated the AI becomes and more about whether customers understand it, control it, and receive a service benefit worth the exchange.