Verified
Starbucks
Experimenting- Coffeehouses
- ~381,000 employees (2025)
- Green Dot Assist on Azure OpenAI
- CTO from Amazon (Jan 2026)
- AI inventory app retired (2026)
Starbucks deploys AI mainly inside its coffeehouses: Green Dot Assist, a generative AI assistant built on Microsoft's Azure OpenAI platform, has answered store teams' questions across North America since Nov 2025, and Smart Queue sequences cafe, drive-thru and mobile orders. Starbucks opened a beta app in ChatGPT in Apr 2026, uses AI-assisted coding to replace vendor software and plans AI ordering and drive-thru tests for fiscal 2027, but it retired an AI inventory counting app in May 2026 and reports no return on AI.
starbucks.comTheir approach to AI
- The platform
- Starbucks' named AI tools are Green Dot Assist, a generative AI assistant on store iPads built with Microsoft, and Smart Queue, an in-house order-sequencing algorithm (Jan 2026); Deep Brew, the Azure-hosted AI platform it described in 2019, did not feature in its 2026 investor materials.
- The models
- Green Dot Assist runs on Microsoft's Azure OpenAI platform, and a beta Starbucks app inside OpenAI's ChatGPT has suggested drinks since Apr 15, 2026; Starbucks has named no other model provider, and Bloomberg did not name the AI coding tools its technology team uses.
- The reach
- Starbucks scaled Green Dot Assist across its North American coffeehouses in Nov 2025 and has published no usage figures; Bloomberg reported in Jul 2026 that its technology team builds software with AI-assisted coding, and Starbucks has named no general AI assistant for support staff.
- The catch
- Starbucks reports no business return for any AI tool and no count of the partners using Green Dot Assist, and it retired its AI inventory counting app in May 2026, about nine months after rollout, after Reuters reported miscounts.
Snapshot
- Employees
- ~381,000
- Sep 28, 2025, before Starbucks China's retail business moved to a joint venture on Mar 30, 2026 (FY2025 Form 10-K)
- Technology budget
- Not disclosed
- About $400M a year on software alone, CTO Anand Varadarajan told staff; enterprise technology on track to cut about $30M in fiscal 2026 (Bloomberg, Jul 2026)
- AI/ML headcount
- Not disclosed
- Bloomberg put the roughly 270 technology roles moving to Nashville at nearly 20% of the technology workforce (2026)
- Who owns AI
- Anand Varadarajan, CTO
- Joined from Amazon on Jan 19, 2026, reporting to Brian Niccol; chief data and analytics officer Bhagyesh Phanse left for Humana in May 2026
- Platform
- Green Dot Assist and Smart Queue
- The assistant was built with Microsoft; the order-sequencing algorithm is proprietary; Deep Brew dates from 2019
- Models
- OpenAI via Microsoft Azure
- Green Dot Assist on Azure OpenAI; a beta Starbucks app in ChatGPT since Apr 15, 2026
- Reach
- North American coffeehouses
- Green Dot Assist on store iPads since Nov 2025; no usage figures published
- Use cases
- A handful in production
- Green Dot Assist, Smart Queue, AI-assisted coding and the ChatGPT app beta; AI inventory counting retired in May 2026
Key findings
Starbucks piloted Green Dot Assist, a generative AI assistant built on Microsoft's Azure OpenAI platform, in 35 stores in Jun 2025 and scaled it across its North American coffeehouses in Nov 2025, but Starbucks has not said how many baristas use it or what it returns.
Starbucks retired Automated Counting, a NomadGo computer vision app that counted coffeehouse inventory on store tablets, in May 2026, about nine months after deploying it across its North American stores, after Reuters reported that it miscounted and mislabeled items.
Anand Varadarajan, who spent nearly two decades at Amazon, became Starbucks' CTO reporting to Brian Niccol on Jan 19, 2026, and Bloomberg reported in Jul 2026 that Starbucks is using AI, including AI-assisted coding, to build in-house replacements for Microsoft and IBM software while spending about $400 million a year on software.
Starbucks launched a beta app inside ChatGPT on Apr 15, 2026 that turns a mood or a photo into a drink suggestion, and Brian Niccol said on Oct 29, 2026 that Starbucks will test background AI models in drive-thrus and work toward conversational and agentic ordering in its app.
Starbucks reports no financial return on AI: its $2 billion cost savings program, more than half delivered by Oct 2026, includes technology-driven efficiencies without separating AI, and Bhagyesh Phanse, the chief data and analytics officer who set out Starbucks' AI approach in Jan 2026, left for Humana in May 2026.
Where Starbucks sits
- Stage
- Experimenting
Starbucks has reported no business return it credits to AI: it gives service times, not returns, for Smart Queue, and its $2 billion cost savings program, more than half delivered by Oct 2026, includes technology-driven efficiencies without separating AI. On rollout, Brian Niccol said Green Dot Assist, a generative AI assistant on Azure OpenAI, was fully scaled across North American coffeehouses in Nov 2025, but Starbucks has published no count of the partners who use it or any other AI tool. Without a return credited to AI or a stated number of users, Starbucks sits at Experimenting, and either one, reported by Starbucks itself, would move it to Piloting. Starbucks' AI ordering companion and drive-thru AI remain plans for fiscal 2027, and Starbucks retired its AI inventory counting app in May 2026 after about nine months.
How the stages are defined - Ownership
- CTO Anand Varadarajan, reporting to the CEO (since Jan 2026)
- Platform
- Hybrid: Green Dot Assist built with Microsoft; Smart Queue and Deep Brew built in-house
- Models
- OpenAI models through Microsoft Azure OpenAI; a beta app in ChatGPT
- Hosting
- Microsoft Azure: Azure OpenAI for Green Dot Assist and Deep Brew's algorithms (2019); the ChatGPT app on OpenAI's platform; no AI data center of its own disclosed
- Reach
- Store teams in North American coffeehouses (Nov 2025); usage not disclosed
How they deploy
Technology in service of the coffeehouse
Starbucks has tied its AI work to the Back to Starbucks turnaround that Brian Niccol began in Sep 2024, which centers on service in the coffeehouse. Niccol said in Dec 2025 that Starbucks will keep technology subordinate to the customer experience and keep baristas making drinks by hand, and Bhagyesh Phanse, then chief data and analytics officer, wrote in Jan 2026 that Starbucks designs AI to strengthen, not replace, the human connection in its coffeehouses.
Tools for store teams first
Starbucks' first generative AI tool is Green Dot Assist, which store teams use on iPads to look up beverage builds, troubleshoot equipment and adjust deployment plans. Brian Niccol told investors in Jan 2026 that Green Dot Assist is also the foundation to test, then develop and scale further AI tools that reduce friction for partners, and Mike Grams said at the Jan 29, 2026 Investor Day that AI could help Starbucks build schedules and forecast peak demand.
Customer ordering through AI
Tressie Lieberman told the Jan 29, 2026 Investor Day that Starbucks is developing an AI ordering companion that suggests a drink from a customer's mood and completes the order. Starbucks opened a beta app in ChatGPT in Apr 2026, and Brian Niccol said on Oct 29, 2026 that Starbucks is working on AI search and discovery in its own app, conversational and agentic ordering over time, and mobile orders placed through emerging AI platforms.
Building more software in-house
Bloomberg reported in Jul 2026 that Starbucks is developing in-house tools with the help of AI to replace a Microsoft system that tracks inventory and an IBM maintenance management tool, with AI-assisted coding key to the IBM replacement, and is building a point-of-sale system to replace Oracle Simphony. CTO Anand Varadarajan told staff there are clear opportunities to reduce the roughly $400 million Starbucks spends on software each year, and in Sep 2026 he named Chennai, India, as the site of a technology hub that will bring work now done by third-party providers in-house. Brian Niccol said fiscal 2027 will be a major technology modernization year, with new inventory ordering, staffing and scheduling, and point-of-sale systems.
Test, then keep or drop
Starbucks says it will test, learn and scale what works, and it has dropped tools that did not. Starbucks retired its Automated Counting app in May 2026, saying it had moved to a single, consistent counting process and was being disciplined about where automation adds value, and Brian Niccol said in Oct 2026 that the drive-thru AI work is just getting started.
- Deployment model
- Starbucks deploys AI through its technology organization, led by CTO Anand Varadarajan, and puts the tools into coffeehouse operations run by COO Mike Grams. Green Dot Assist sits on store iPads in North American coffeehouses, Smart Queue runs behind the bar as part of the Green Apron Service operating model, and digital and loyalty teams own customer-facing AI such as the ChatGPT app. Starbucks has named no general AI assistant for its support center staff, though Bloomberg reported in Jul 2026 that its technology team uses AI-assisted coding, and its international markets, nearly 90% licensed after the China joint venture closed on Mar 30, 2026, have not been named as users of Green Dot Assist.
- Model providers
- Starbucks built Green Dot Assist with Microsoft on the Azure OpenAI platform, and Deb Hall Lefevre said in Jun 2025 that the Microsoft partnership includes a grounding engine to keep answers accurate. Starbucks' customer app inside ChatGPT runs on OpenAI's platform, where users tag Starbucks and then finish checkout in the Starbucks app or on Starbucks.com. Smart Queue is a proprietary algorithm, and Starbucks has not named the models behind its planned ordering companion, the background AI models it will test in drive-thrus, or the AI coding tools its technology team uses. Automated Counting, retired in May 2026, used NomadGo's computer vision, LiDAR and camera data.
- Data and infrastructure
- Starbucks runs its named AI on Microsoft Azure: Green Dot Assist uses Azure OpenAI, Restaurant Dive reported in Nov 2019 that Deep Brew's algorithms sit on Azure, and Microsoft said in May 2019 that Starbucks' app recommendations came from a reinforcement learning platform built and hosted on Azure. Starbucks' FY2025 Form 10-K says it depends on third-party providers for key systems, some without full redundancy, and does not describe its own data centers. Tressie Lieberman said in Jan 2026 that the Starbucks app and CRM are not yet highly personalized, and Brian Niccol said in Oct 2026 that Starbucks will build more personalized recommendations into the app in fiscal 2027.
Platforms and internal tools
Green Dot AssistJun 2025 pilot; Nov 2025 across North America
Green Dot Assist is Starbucks' generative AI assistant for coffeehouse teams: baristas and managers type or speak questions on an iPad behind the counter and get conversational answers on recipes, routines, service standards and equipment errors. Starbucks showed it to more than 14,000 North American store managers in Las Vegas in Jun 2025, piloted it in 35 stores and scaled it across North American coffeehouses in Nov 2025. Deb Hall Lefevre said later versions could open IT tickets for equipment issues and suggest substitutes when a barista calls out.
- Models
- Microsoft Azure OpenAI
Smart QueueAug 2025 across U.S. company-operated coffeehouses
Smart Queue is Starbucks' proprietary order-sequencing algorithm, which orders drink production across cafe, drive-thru, mobile and delivery so mobile orders stop crowding out customers in the cafe. Starbucks piloted an in-store prioritization algorithm in three stores in early 2025, rolled Smart Queue out with Green Apron Service from mid-Aug 2025, and counted it among its AI tools in Jan 2026.
Starbucks app in ChatGPT (beta)Apr 15, 2026
The Starbucks app in ChatGPT lets customers tag Starbucks, describe a mood or upload a photo, get a drink suggestion, customize it and pick a coffeehouse, then finish the order in the Starbucks app or on Starbucks.com.
- Models
- OpenAI ChatGPT
Automated Counting (retired)Sep 2025; retired May 2026
Automated Counting was a NomadGo app on store tablets that used computer vision, 3D spatial intelligence and augmented reality to count inventory in fridges and stockrooms. Starbucks said in Sep 2025 that the app was live in thousands of coffeehouses and would reach every North American company-operated coffeehouse by the end of that month, and it retired the app in May 2026 after Reuters reported that it confused product names or missed items.
Deep Brew2017 project; described publicly in 2019
Deep Brew is the AI initiative Starbucks said in Oct 2019 would drive its app personalization engine, in-store inventory management and labor deployment, with its algorithms on Microsoft Azure; Microsoft said in May 2019 that a reinforcement learning platform built and hosted on Azure tailored app order suggestions for 16 million active Starbucks Rewards members. Forbes Argentina called DeepBrew Starbucks' own AI platform in a Sep 2025 interview with Bhagyesh Phanse, but Starbucks did not name Deep Brew at its Jan 2026 Investor Day or on its fiscal 2026 earnings calls.
Vendors and partners
MicrosoftBy 2019
Azure OpenAI platform behind Green Dot Assist; Azure cloud for Deep Brew and earlier personalization work · Microsoft hosted Starbucks' reinforcement learning recommendation platform on Azure by May 2019, and Green Dot Assist, built with Microsoft on Azure OpenAI, launched in Jun 2025. Bloomberg reported in Jul 2026 that Starbucks is building a replacement for a Microsoft system that tracks inventory.
OpenAI2025
Models behind Green Dot Assist through Azure; host of the Starbucks app in ChatGPT · Starbucks launched its beta app in ChatGPT on Apr 15, 2026, with checkout kept in the Starbucks app or on Starbucks.com.
NomadGoSep 2025
Computer vision inventory counting app (retired) · Starbucks ended its use of NomadGo's app in May 2026; NomadGo said it keeps learning from customer and user feedback.
Oracle
Simphony point-of-sale system · Bloomberg reported in Jul 2026 that Starbucks has worked for several years on a point-of-sale system of its own to replace Oracle Simphony; Starbucks plans new point-of-sale systems in fiscal 2027.
IBM
Maintenance management tool · Bloomberg reported in Jul 2026 that Starbucks used AI-assisted coding to develop an alternative to an IBM maintenance management tool.
Leadership
AI, data and platform leaders
Executives
Board
AI org map
CTO-led technology organization; AI tools put into coffeehouse operations and the customer app
Starbucks has no chief AI officer. CTO Anand Varadarajan, hired from Amazon in Jan 2026, runs technology and reports to Brian Niccol; the data and analytics group has had no publicly named chief since Bhagyesh Phanse left for Humana in May 2026; and the AI tools land in coffeehouse operations under COO Mike Grams and in the app under the brand, digital and loyalty teams. None of Starbucks' technology or data executives was among the six executive officers in its FY2025 Form 10-K (Nov 2025).
Starbucks has not disclosed how many data scientists or AI engineers it employs. Starbucks' technology organization, led by CTO Anand Varadarajan, has teams in Seattle, Nashville, London and Hong Kong; Bloomberg reported in 2026 that about 270 technology roles, nearly 20% of the technology workforce, would move from Seattle to Nashville, and Reuters reported in Sep 2026 that a new Chennai technology hub would create 800 technology jobs. Starbucks' data and analytics group, which develops AI and machine learning models, lost its chief, Bhagyesh Phanse, to Humana in May 2026. Starbucks employed about 9,000 U.S. staff outside its stores in corporate support, store development, roasting, manufacturing, warehousing and distribution, plus about 5,000 in regional support abroad, at Sep 28, 2025, and has cut about 2,300 support roles in three rounds since Feb 2025.
Reporting lines
Chairman and CEO
Chief Financial Officer
Technology
Anand Varadarajan, EVP and CTO
Reports directly to Brian Niccol (Dec 2025 appointment); Deb Hall Lefevre, the previous CTO, reported to CFO Cathy Smith from Jun 2025
Chief Operating Officer
Coffeehouse operations, coffeehouse development and the global supply chain (Jun 2025); runs the stores where Green Dot Assist and Smart Queue operate
Global Brand
Tressie Lieberman, Global Chief Brand Officer
Line to the CEO inferred, not published; announced the AI ordering companion
Data and Analytics
Bhagyesh Phanse led it as SVP and chief data and analytics officer from Apr 2025 until he left for Humana in May 2026; no successor or reporting line published
Law and Corporate Affairs
Pilar Ramos, Chief Legal Officer
Co-manages the Risk Management Committee with the CFO; reporting line not published
Starbucks publishes no org chart. Brian Niccol moved CTO Deb Hall Lefevre to report to CFO Cathy Smith in Jun 2025, and Nation's Restaurant News reported in Dec 2025 that Anand Varadarajan, who replaced her in Jan 2026, reports directly to Niccol. Starbucks has not said who leads data and analytics since Bhagyesh Phanse left in May 2026, nor where that group or Paul Riedel's digital and loyalty team sit. The FY2025 Form 10-K says an unnamed senior vice president and chief information security officer runs the cybersecurity program and briefs the board's Audit Committee each quarter.
Named people
Use cases
Starbucks has named a handful of AI use cases: Green Dot Assist for store teams and Smart Queue order sequencing in production, AI-assisted coding for in-house software, a beta app in ChatGPT, and plans for an AI ordering companion, drive-thru AI and AI-assisted scheduling and replenishment. Starbucks has published no business return for any of them; the figures it gives are service times, such as more than 80% of U.S. company-operated coffeehouses averaging four minutes or less in the cafe after Smart Queue (Oct 2025). Starbucks reports no firm-level return on AI, and its $2 billion cost savings program does not separate AI from other technology and procurement savings.
| Use case | Area | Status | ROI |
|---|---|---|---|
| Green Dot Assist Green Dot Assist gives baristas and managers conversational answers on an in-store iPad about drink recipes, equipment troubleshooting, routines and deployment plans, with a Microsoft grounding engine meant to keep answers accurate. | Coffeehouse operations | In production | |
| Smart Queue order sequencing Smart Queue sequences drink production across cafe, drive-thru, mobile and delivery orders as part of the Green Apron Service model, and Starbucks says it keeps refining the algorithm as transactions grow. | Coffeehouse operations | In production | |
| AI-assisted coding for in-house software Starbucks' technology team is building in-house replacements for a Microsoft inventory system and an IBM maintenance management tool with the help of AI; Bloomberg reported that AI-assisted coding was key to the IBM replacement, which could roll out by the end of 2027 pending tests. | Technology | Pilot | |
| Starbucks app in ChatGPT Customers describe a mood or upload a photo in ChatGPT, get a drink suggestion, customize it, choose a coffeehouse and finish the order in the Starbucks app or on Starbucks.com. | Digital | Pilot | |
| AI ordering companion and app search Starbucks is developing an AI ordering companion that turns quick prompts into a drink and an order, and plans AI search and discovery in its app with conversational and agentic ordering over time. | Digital | Announced | |
| Personalized app recommendations Starbucks plans more personalized menu content and recommendations in its app and loyalty program in fiscal 2027; Microsoft described a Starbucks reinforcement learning recommendation platform on Azure in May 2019. | Digital | Announced | |
| Background AI order entry Starbucks will test background AI models in fiscal 2027 that listen to drive-thru orders and enter them while the barista keeps taking the order in person. | Drive-thru | Announced | |
| Demand forecasting and replenishment Starbucks is testing changes to demand forecasting, ordering accuracy and replenishment cycles (Oct 2026), work Mike Grams described in Jan 2026 as AI-ready platforms aimed at daily replenishment for 90% of U.S. coffeehouses. | Supply chain | Pilot | |
| Staffing and scheduling Mike Grams said in Jan 2026 that AI could help Starbucks build schedules that balance partners' preferred hours with forecast peak demand and weather; Starbucks will roll out new staffing and scheduling systems in fiscal 2027 without saying whether they use AI. | Coffeehouse operations | Announced |
Adoption
Starbucks used machine learning for app recommendations by 2019, when Microsoft said a reinforcement learning platform on Azure served 16 million active Starbucks Rewards members. Starbucks' generative AI rollout began with Green Dot Assist in 35 stores in Jun 2025 and reached North American coffeehouses in Nov 2025, while Smart Queue spread with Green Apron Service across U.S. company-operated coffeehouses from Aug 2025. Starbucks has published no count of active Green Dot Assist users and no adoption curve, has named no general AI assistant for support staff beyond the AI-assisted coding Bloomberg reported in its technology team in Jul 2026, and retired one store-level AI tool, Automated Counting, in May 2026.
By the numbers
Rollout, first pilot to today
May 2019
Microsoft said Starbucks used a reinforcement learning platform built and hosted on Azure to personalize app order suggestions for 16 million active Starbucks Rewards members.
Oct 2019
Starbucks said on its fiscal 2019 fourth-quarter earnings call that its Deep Brew AI initiative would drive its personalization engine and in-store inventory management.
Jan 2025
Brian Niccol said Starbucks was piloting an in-store order prioritization algorithm in three stores to stop mobile orders from crowding out cafe customers.
Jun 2025
Starbucks showed Green Dot Assist, built on Microsoft's Azure OpenAI platform, to more than 14,000 North American store managers and began a 35-store pilot.
Aug 2025
Starbucks began rolling out Green Apron Service, including Smart Queue, across U.S. company-operated coffeehouses in mid-Aug 2025.
Sep 2025
Starbucks deployed Automated Counting, a NomadGo computer vision app, in thousands of coffeehouses and said every North American company-operated coffeehouse would have it by the end of the month.
Nov 2025
Starbucks fully scaled Green Dot Assist across its North American coffeehouses.
Jan 2026
Starbucks announced an AI ordering companion and a drive-thru AI pilot at its Jan 29 Investor Day, ten days after CTO Anand Varadarajan started.
Apr 2026
Starbucks launched a beta Starbucks app in ChatGPT for drink discovery and order starts.
May 2026
Starbucks ended its use of Automated Counting across North America after about nine months.
Jul 2026
Bloomberg reported that Starbucks' technology team was using AI-assisted coding to build in-house replacements for Microsoft and IBM software.
Oct 2026
Brian Niccol set fiscal 2027 plans for AI search and discovery in the Starbucks app and tests of background AI models in drive-thrus.
Economics & controls
Spend and ROI
Technology budget
Starbucks does not disclose a technology or AI budget. Bloomberg reported in Jul 2026 that CTO Anand Varadarajan told employees Starbucks spends roughly $400 million a year on software alone and sees clear opportunities to reduce it, and that the enterprise technology team was on track to cut its budget by about $30 million in the fiscal year ending Sep 2026, including about $10 million of software and $13 million mostly from fewer professional services contractors. Starbucks guided fiscal 2027 capital expenditures to about 5% of net revenues, mainly for coffeehouse uplifts and strategic priorities, and called fiscal 2027 a major technology modernization year.
What AI itself costs
Starbucks has not disclosed what Green Dot Assist, the ChatGPT app, its AI coding tools or the retired Automated Counting app cost, or what it pays Microsoft, OpenAI or NomadGo.
Reported ROI
Starbucks reports no financial return on AI. Brian Niccol said at the Jan 2026 Investor Day that Starbucks had identified about $2 billion of cost savings over two years and that more efficiency could come from how it uses technology and AI, and CFO Cathy Smith said Starbucks will offset general and administrative inflation with technology-driven productivity where practical. Starbucks said in Oct 2026 that it had achieved more than half of the $2 billion gross savings goal, without saying how much came from AI.
Productivity markers
Starbucks reports service times, not returns, for its AI tools: Brian Niccol said in Jul 2025 that coffeehouses with Smart Queue saw a double-digit improvement in cafe orders handed off in under four minutes, and in Oct 2025 that more than 80% of U.S. company-operated coffeehouses averaged four minutes or less in the cafe since Smart Queue launched. Starbucks said in Sep 2025 that Automated Counting let stores count inventory eight times more often, then retired the app in May 2026.
Headcount
Starbucks employed about 381,000 people at Sep 28, 2025, about 367,000 of them in company-operated stores. Starbucks cut 1,100 corporate support roles in Feb 2025, 900 in Sep 2025 and about 300 U.S. support roles in May 2026, and has not credited any of the cuts to AI. Brian Niccol said in Apr 2026 that Starbucks does not see its way forward with cutting store labor and will use technology, equipment and processes to get more throughput from the hours it has.
Governance and risk
Board and risk oversight
Starbucks' board oversees technology risk through its Audit and Compliance Committee, which covers cybersecurity, data privacy and technology risks and gets quarterly updates from the chief information security officer, according to the 2026 proxy statement. A management Risk Management Committee, co-managed by CFO Cathy Smith and chief legal officer Pilar Ramos, meets at least quarterly, according to the FY2025 Form 10-K. Starbucks has not disclosed an AI policy, an AI council or board oversight specific to AI.
Disclosed AI risks
Starbucks' FY2025 Form 10-K says AI and machine learning may not deliver expected efficiencies and could bring risks of cybersecurity, data privacy, inaccuracies, hallucinations, bias and intellectual property infringement that grow as reliance on them grows. The 10-K also says AI regulation is uncertain and could add compliance costs or limit how Starbucks uses the technology, and that privacy authorities are setting new rules on processing personal data with AI.
Accuracy and grounding
Deb Hall Lefevre said in Jun 2025 that Starbucks' partnership with Microsoft includes a grounding engine to keep Green Dot Assist's answers accurate. Starbucks retired Automated Counting in May 2026 after Reuters reported miscounts and mislabeled items, and shared employee comments thanking the company for trusting partners over the app's counts.
People stay in the order
Brian Niccol said in Dec 2025 that robots making espresso drinks would take the soul out of the experience, and in Jan 2026 that the drive-thru AI pilot will listen in the background while baristas keep taking orders in person. Bhagyesh Phanse, then chief data and analytics officer, wrote in Jan 2026 that Starbucks designs AI to strengthen, not replace, the human connection.
Customer data in third-party AI
Starbucks keeps checkout for orders started in ChatGPT inside the Starbucks app or on Starbucks.com, and Paul Riedel called the ChatGPT app a beta to listen, learn and refine. Starbucks has not said what customer data the ChatGPT app shares with OpenAI.
Recent signals
Oct 29, 2026
Brian Niccol said Starbucks will test background AI models in drive-thrus in fiscal 2027, work on AI search and discovery in its app with conversational and agentic ordering over time, and roll out new staffing, scheduling and point-of-sale systems.
Sep 21, 2026
CTO Anand Varadarajan named Chennai, India, as the site of a Starbucks technology hub opening in fiscal 2027 to bring work now done by third-party service providers in-house over time.
Jul 29, 2026
Brian Niccol said Starbucks keeps making Smart Queue smarter and that fiscal 2027 will be a technology modernization year with new inventory ordering, staffing and scheduling, and point-of-sale systems.
Jul 9, 2026
Bloomberg reported that Starbucks is using AI, including AI-assisted coding, to build in-house replacements for a Microsoft inventory system and an IBM maintenance tool, and that CTO Anand Varadarajan put annual software spend at roughly $400 million.
May 21, 2026
Reuters reported that Starbucks had retired the Automated Counting app across North America after it miscounted and mislabeled items; Starbucks said it wanted to standardize how inventory is counted.
May 2026
Bhagyesh Phanse, Starbucks' SVP and chief data and analytics officer, left to become Humana's chief data and analytics officer; Starbucks has not publicly named a successor.
Apr 15, 2026
Starbucks launched a beta app in ChatGPT that suggests drinks from a mood or a photo and hands the order to the Starbucks app or Starbucks.com.
Feb 2026
Reuters reported that Starbucks employees and managers said the Automated Counting app often made errors, while Starbucks said the app had improved product availability.
Jan 29, 2026
Starbucks' Investor Day announced an AI ordering companion, a drive-thru pilot in which AI enters orders in the background, and AI-ready supply chain platforms aimed at daily replenishment.
Jan 28, 2026
Brian Niccol said Green Dot Assist had been fully scaled across North American coffeehouses in Nov 2025 and introduced Anand Varadarajan as CTO.
Dec 19, 2025
Starbucks named Anand Varadarajan, an Amazon grocery technology leader, as CTO from Jan 19, 2026, reporting to Brian Niccol and replacing Deb Hall Lefevre, who retired in Sep 2025.
Dec 9, 2025
Brian Niccol told the Wall Street Journal's CEO Council that Starbucks will keep technology subordinate to the customer experience and does not want robots making its drinks.
Quick answers
- How does Starbucks deploy AI?
- Starbucks deploys AI mainly inside its coffeehouses: Green Dot Assist, a generative AI assistant built on Microsoft's Azure OpenAI platform, has answered store teams' questions across North America since Nov 2025, and Smart Queue sequences cafe, drive-thru and mobile orders. Starbucks opened a beta app in ChatGPT in Apr 2026, uses AI-assisted coding to replace vendor software and plans AI ordering and drive-thru tests for fiscal 2027, but it retired an AI inventory counting app in May 2026 and reports no return on AI.
- What does Starbucks use AI for?
- Starbucks has named a handful of AI use cases: Green Dot Assist for store teams and Smart Queue order sequencing in production, AI-assisted coding for in-house software, a beta app in ChatGPT, and plans for an AI ordering companion, drive-thru AI and AI-assisted scheduling and replenishment. Starbucks has published no business return for any of them; the figures it gives are service times, such as more than 80% of U.S. company-operated coffeehouses averaging four minutes or less in the cafe after Smart Queue (Oct 2025). Starbucks reports no firm-level return on AI, and its $2 billion cost savings program does not separate AI from other technology and procurement savings.
- Who leads AI at Starbucks?
- AI at Starbucks is led by Anand Varadarajan (Executive Vice President and Chief Technology Officer); Paul Riedel (Senior Vice President, Digital and Loyalty). Ownership of AI on the org chart: CTO Anand Varadarajan, reporting to the CEO (since Jan 2026).
- What AI platform does Starbucks use?
- Starbucks' named AI tools are Green Dot Assist, a generative AI assistant on store iPads built with Microsoft, and Smart Queue, an in-house order-sequencing algorithm (Jan 2026); Deep Brew, the Azure-hosted AI platform it described in 2019, did not feature in its 2026 investor materials.
- Which AI models does Starbucks use?
- Green Dot Assist runs on Microsoft's Azure OpenAI platform, and a beta Starbucks app inside OpenAI's ChatGPT has suggested drinks since Apr 15, 2026; Starbucks has named no other model provider, and Bloomberg did not name the AI coding tools its technology team uses.
- How many Starbucks employees use AI?
- Starbucks scaled Green Dot Assist across its North American coffeehouses in Nov 2025 and has published no usage figures; Bloomberg reported in Jul 2026 that its technology team builds software with AI-assisted coding, and Starbucks has named no general AI assistant for support staff.
- How far has AI spread at Starbucks?
- Starbucks has reported no business return it credits to AI: it gives service times, not returns, for Smart Queue, and its $2 billion cost savings program, more than half delivered by Oct 2026, includes technology-driven efficiencies without separating AI. On rollout, Brian Niccol said Green Dot Assist, a generative AI assistant on Azure OpenAI, was fully scaled across North American coffeehouses in Nov 2025, but Starbucks has published no count of the partners who use it or any other AI tool. Without a return credited to AI or a stated number of users, Starbucks sits at Experimenting, and either one, reported by Starbucks itself, would move it to Piloting. Starbucks' AI ordering companion and drive-thru AI remain plans for fiscal 2027, and Starbucks retired its AI inventory counting app in May 2026 after about nine months.
Open questions
- How many Starbucks baristas and managers use Green Dot Assist each week, and what does it cost to run?
- What business return does Starbucks credit to Green Dot Assist and Smart Queue, beyond service times?
- Which OpenAI models run behind Green Dot Assist, and is any Starbucks data used to tune them?
- Who has led Starbucks' data and analytics group since Bhagyesh Phanse left for Humana in May 2026, and where do that group and Paul Riedel's digital team report?
- Does Starbucks give its support center staff a general generative AI assistant, and which AI coding tools and models does its technology team use?
- How much of Starbucks' $2 billion cost savings program comes from AI rather than procurement, store builds and the support center reorganization?
- How much will the in-house replacements for Microsoft, IBM and Oracle software, built partly with AI-assisted coding, save against the roughly $400 million software bill?
- What did Automated Counting cost, what replaced it, and how will Starbucks judge the AI supply chain and scheduling tools it plans for fiscal 2027?
- Does Deep Brew still run, given that Tressie Lieberman said in Jan 2026 that the app is not yet highly personalized?
- Which model provider will power the AI ordering companion and the drive-thru AI tests, and what customer data does the ChatGPT app share with OpenAI?
- How many employees does Starbucks have after its China retail business moved to a joint venture on Mar 30, 2026, a figure due in the FY2026 Form 10-K?
Sources
Starbucks
Filings
- Starbucks FY2025 Form 10-K (filed Nov 14, 2025)
- Starbucks 2026 proxy statement (Jan 26, 2026)
- Q4 and full fiscal year 2026 earnings release, 8-K exhibit (Oct 29, 2026)
- Q3 FY2026 Form 10-Q (Jul 29, 2026)
- 8-K on support organization restructuring (May 2026)
- 8-K on North America coffeehouse closures (Sep 2026)
Earnings call transcripts
- Q4 FY2026 earnings call, The Motley Fool (Oct 29, 2026)
- Q3 FY2026 earnings call, The Motley Fool (Jul 29, 2026)
- Q2 FY2026 earnings call, The Motley Fool (Apr 28, 2026)
- Q1 FY2026 earnings call, The Motley Fool (Jan 28, 2026)
- Q4 FY2025 earnings call, The Motley Fool (Oct 29, 2025)
- Q3 FY2025 earnings call, The Motley Fool (Jul 29, 2025)
- Q1 FY2025 earnings call, The Motley Fool (Jan 28, 2025)
News
- Restaurant Dive: Starbucks' AI project is turning past investments into data insights (Nov 1, 2019)
- CNBC: Starbucks to roll out Microsoft Azure OpenAI assistant for baristas (Jun 10, 2025)
- Restaurant Dive: Starbucks deploys virtual assistant to improve workflows (Jun 11, 2025)
- Restaurant Dive: Starbucks revives COO position (Jun 5, 2025)
- CDO Magazine: Starbucks promotes Bhagyesh Phanse as chief data and analytics officer (Apr 16, 2025)
- Restaurant Dive: Starbucks deploys AI for in-store inventory management (Sep 3, 2025)
- Forbes Argentina: interview with Bhagyesh Phanse on data and AI at Starbucks (Sep 11, 2025)
- Restaurant Dive: Starbucks' on-premise experience could stave off coffee competitors (Dec 10, 2025)
- Nation's Restaurant News: Starbucks names former Amazon exec Anand Varadarajan chief technology officer (Dec 22, 2025)
- Restaurant Dive: Starbucks hires technology chief from Amazon (Dec 22, 2025)
- PYMNTS: Starbucks says digital platforms are key piece of turnaround strategy (Jan 28, 2026)
- Restaurant Dive: 3 takeaways from Starbucks' Investor Day (Jan 30, 2026)
- eWeek: Starbucks developing AI ordering companion (Jan 30, 2026)
- Restaurant Dive: Starbucks lets customers discover drinks, start orders in ChatGPT (Apr 15, 2026)
- PYMNTS: Starbucks brews AI-powered drink picks inside ChatGPT (Apr 15, 2026)
- Business Insider: Starbucks is letting ChatGPT decide what you drink (Apr 15, 2026)
- Restaurant Dive: Starbucks Nashville HQ will cost $100M, employ 2K workers (Apr 22, 2026)
- Restaurant Dive: Starbucks to cut 300 jobs, close 4 support centers (May 15, 2026)
- Business Insider: Starbucks announces third round of corporate layoffs (May 15, 2026)
- People Matters: Starbucks opens India tech hub while cutting jobs, citing Bloomberg (May 18, 2026)
- CNBC: Starbucks scraps AI inventory tool across North America, citing Reuters (May 21, 2026)
- Restaurant Dive: Starbucks ditches AI inventory system after just 9 months (May 22, 2026)
- PYMNTS: Starbucks scraps AI inventory tool after counting errors (May 22, 2026)
- CDO Magazine: Humana names Bhagyesh Phanse SVP, chief data and analytics officer (May 27, 2026)
- Fortune, from Bloomberg: Starbucks to use AI to replace Microsoft, IBM software (Jul 9, 2026)
- PYMNTS: Starbucks brews in-house tools in shift from Big Tech, citing Bloomberg (Jul 9, 2026)
- Restaurant Dive: Starbucks targets 24-hour inventory replenishment (Aug 7, 2026)
- Reuters via TimesLIVE: Starbucks plans global tech centre in India (Sep 21, 2026)
- Pulse 2.0: Starbucks to open Chennai technology hub in FY27 (Sep 21, 2026)
- CX Dive: Starbucks puts greater personalization on the menu for 2027 (Oct 30, 2026)
- CX Dive: Starbucks CEO zeroes in on bringing order to mobile ordering (Jan 31, 2025)
Data confidence
- Confirmed
- Stated in a primary or reputable public source.
- Estimated
- Derived or approximated from public reporting.
- Unconfirmed
- Reported, but not yet corroborated.
Starbucks profile · Enterprise AI diffusion report by Imagine · facts checked against their sources on Nov 1, 2026
