Starbucks has launched a new beta app inside ChatGPT that turns your mood, vibe, or outfit into a personalized drink recommendation, instead of asking you to start with a traditional menu. The feature marks one of Starbucks most visible public uses of generative AI for customers, building on earlier AI tools that focused mainly on baristas and store operations.
Inside ChatGPT, users can open the Starbucks app and describe what they are looking for in natural language, such as a sugar free drink, something that captures a sunset feeling, or a coffee that matches their workday vibe. The chatbot then suggests a drink, complete with customization options like cold foam or matcha powder, so the recommendation feels tailored instead of generic. Customers can also specify preferences around protein content or sweetness level to steer the result in a more health focused or indulgent direction.
Once a user settles on a drink, the experience does not end inside ChatGPT, they can choose a store location and then complete the order through the Starbucks app or website. This keeps payment and fulfillment within Starbucks existing digital ecosystem, while using ChatGPT as a front end discovery and inspiration layer. Starbucks executives describe this beta as an opportunity to listen and learn how customers naturally talk about their cravings and feelings when they are not constrained by a rigid menu format.
Starbucks says it is responding to a clear pattern in how people now make choices, with many customers starting with a feeling rather than a specific menu item. Paul Riedel, senior vice president of digital and loyalty, said that over the past year the company has seen that customers often arrive with a mood or vibe first, and only later translate that into a drink. The ChatGPT app is designed to meet them at that moment of inspiration and quickly map soft prompts like “comforting,” “energizing,” or “something that goes with my outfit” into a concrete beverage order.
This approach also aligns with broader shifts in how people use AI, with ChatGPT already acting as a meal planner, personal stylist, and shopping assistant for many users. Starbucks is effectively inserting its brand into that pattern by letting the same AI that helps plan a day or outfit also suggest a matching coffee or tea. The company hopes that turning drink discovery into a conversational, playful interaction will make customers more excited about what they are ordering and more likely to try something new.
The ChatGPT beta is one visible piece of a larger AI strategy that Starbucks has been rolling out across its business over the past few years. On the operations side, the chain is piloting and expanding Green Dot Assist, a generative AI virtual assistant that runs on in store devices to help baristas with recipe reminders, ingredient substitutions, equipment troubleshooting, and even shift backfilling support. Green Dot Assist is built on OpenAI technology via Microsoft Azure and is being piloted in dozens of locations before a wider rollout across the United States and Canada.
Alongside that, Starbucks has invested in data driven personalization through its Deep Brew and related AI systems, which power more than four hundred thousand unique experiences daily across its app and loyalty program according to executive commentary. These systems analyze purchase history, time of day, weather, and other context to recommend drinks and food that are more likely to resonate with each individual customer. The new ChatGPT experience complements these efforts by collecting richer language about how people describe their moods and preferences, which can feed back into Starbucks own recommendation engines over time.
For Starbucks, putting a branded experience inside ChatGPT achieves several goals at once. It reaches a tech savvy audience where they already spend time, positions Starbucks as an early mover on consumer facing generative AI, and creates a playful on ramp into its core ordering channels. Instead of forcing customers to open the Starbucks app first, the brand can meet users mid conversation as they are planning their day or looking for ideas.
At the same time, this experiment gives Starbucks valuable insight into the language customers actually use when they are not constrained by menu categories or traditional search filters. Those insights can help refine both AI prompts and product positioning across channels, from the main Starbucks app to in store campaigns and loyalty messaging. As Riedel put it, the beta is a test bed where the company can listen, learn, and refine, while continuing to explore new ways technology can delight customers and support baristas.
Starbucks describes this ChatGPT integration as “only the beginning” of how it plans to use conversational AI in customer journeys. The company is already developing its own AI ordering companion for the Starbucks app, an assistant that lets customers describe cravings in natural language and then handles the translation to drinks behind the scenes. Paired with tools like Green Dot Assist behind the counter and Deep Brew driven personalization in the app, the ChatGPT beta shows how Starbucks is trying to connect discovery, recommendation, and execution into one continuous AI powered loop.
For now, the success of the beta will be measured less by immediate sales and more by engagement, customer feedback, and the quality of prompts and data it generates. If customers embrace feelings based ordering and AI curated drinks, Starbucks could be setting a template for how other restaurant and retail brands use platforms like ChatGPT as discovery engines that feed into their own apps and stores.
Be the first to post comment!
