Starbucks tests an AI assistant to help baristas make drinks faster

Peak coffee rush hours are chaotic: baristas scrambling to remember complex custom recipes while espresso machines spit out cryptic error codes.

Now, Starbucks wants OpenAI to work the bar. In an announcement on Tuesday, Chief Technology Officer Deb Hall Lefevre revealed that the chain is currently testing an internal virtual assistant dubbed Green Dot Assist across 35 stores.

Built on Microsoft Azure's OpenAI platform, the assistant is slated to expand to all company-operated stores across the U.S. and Canada during Starbucks' fiscal 2026 starting this fall, according to CNBC.

An on-demand coach inside the store iPad

Instead of a customer-facing bot, Green Dot Assist acts as an always-on mentor for store staff via in-store iPads.

Employees can ask plain-language questions and get instant conversational guidance. The assistant helps crew members double-check specific drink ingredients, teaches new hires standard recipes on the fly, and walks staff through troubleshooting broken machines.

Store managers can also use it behind the scenes, such as quickly surfacing available employees when someone calls out sick at the last minute.

Alongside the assistant, Starbucks is working on a next-generation point-of-sale system and a mobile-order sequence algorithm to smooth out pickup jams. The company aims to hand over in-person orders within 4 minutes and mobile orders within 12 to 15 minutes.

Fueling the turnaround plan

These technical upgrades fall directly under the chain's "Back to Starbucks" plan, designed to reduce operational friction and let baristas focus on human connection.

Starbucks is fighting through five straight quarters of sluggish same-store sales and declining foot traffic. Since taking over as CEO last September, Brian Niccol has executed aggressive operational shifts.

Niccol has already revived the self-serve condiment bars, reversed the controversial open-door policy, slashed unpopular menu items, restructured corporate roles, and brought back the Chief Operating Officer position originally eliminated in 2022.

Equipping stores with generative AI is another piece of the puzzle: streamlining back-of-house bottlenecks so baristas can actually spend time crafting beverages and serving people.

Why it matters

We think Starbucks is taking a genuinely smart approach here by avoiding the trap of gimmicky AI cashiers.

Fast-food drive-thrus that rolled out public-facing AI order bots regularly faced customer backlash over misheard orders and frustrating conversational loops. Starbucks is deliberately keeping the AI invisible behind the counter, targeting high turnover and cumbersome store training instead.

For frontline workers, memorizing dozens of customized seasonal recipes is notoriously stressful. Giving staff an instant cheat sheet on an iPad should materially reduce drink remake waste and shorten line bottlenecks.

That said, high-volume retail has unique pitfalls. When lines are out the door, pausing to tap on a screen or read AI outputs is the last thing a rushed barista wants to do. If the UX isn't lightning-fast, workers will instantly abandon the tool and just holler across the counter for a manager.

For the broader service industry, Starbucks is showing that the most immediate ROI for enterprise generative AI might not be replacing human labor, but supercharging backline operational knowledge.