๐ฐThe VergeโขStalecollected in 1m
McDonald's AI drive-thru strategy and its industry implications
๐กUnderstand how enterprise-scale voice AI is being deployed in real-world, high-noise retail environments.
โก 30-Second TL;DR
What Changed
McDonald's began testing AI voice-ordering at 10 Chicago locations in 2021.
Why It Matters
The widespread adoption of voice-ordering AI in fast-food indicates a shift toward automated service environments. It highlights the potential for specialized conversational models to replace human labor in high-volume, repetitive tasks.
What To Do Next
Analyze the latency and accuracy requirements for edge-deployed voice AI in high-noise environments to improve your own conversational agent performance.
Who should care:Enterprise & Security Teams
Key Points
- โขMcDonald's began testing AI voice-ordering at 10 Chicago locations in 2021.
- โขThe initiative was powered by technology from the startup Apprente, acquired in 2019.
- โขFast-food chains are increasingly adopting conversational AI to automate customer interactions.
๐ง Deep Insight
Web-grounded analysis with 27 cited sources.
๐ Enhanced Key Takeaways
- โขMcDonald's initially acquired Apprente in 2019 to establish McD Tech Labs, aiming to develop voice-based conversational technology for drive-thrus, mobile ordering, and kiosks.
- โขIn October 2021, McDonald's sold McD Tech Labs to IBM, entering a partnership to accelerate the development and deployment of its Automated Order Taking (AOT) technology.
- โขMcDonald's ended its global partnership with IBM on AI drive-thru technology in June 2024, after deploying the system in over 100 locations, citing a desire to explore broader voice ordering solutions.
- โขThe IBM-powered AI system achieved an accuracy rate in the low 80% range (approximately 80-85%), which fell short of McDonald's target of 95%+ for widespread adoption and led to several viral social media incidents of incorrect orders.
- โขMcDonald's is now collaborating with Google Cloud, a partnership initiated in 2023, to integrate new AI technologies, including voice-activated AI chatbots for drive-thrus and AI-powered accuracy scales to verify order contents.
๐ Competitor Analysisโธ Show
| Chain/Vendor | AI Solution | Key Features/Partnership | Reported Accuracy/Benchmarks |
|---|---|---|---|
| McDonald's | Automated Order Taking (AOT) (IBM) | Voice ordering, multi-item conversational | Low 80-85% accuracy (IBM system) |
| McDonald's | AI Voice Chatbots (Google Cloud) | Voice ordering, new platform 'Edge', AI accuracy scales | Aiming for improved accuracy |
| Checkers & Rally's | Presto Voice / Valyant AI | Drive-thru automation, custom voices, cross-selling | Presto claims 95% accuracy, 20-second throughput improvement, 9 hours/day labor reduction |
| White Castle | SoundHound | Voice AI in drive-thrus | Deployed in ~100 drive-thrus (30% of footprint) |
| Wendy's | FreshAI (Google Cloud) | AI chatbot for drive-thru employees | Expanding to 500-600 locations by end of 2025 |
| Taco Bell (Yum Brands) | Nvidia AI agents | AI agents for drive-thrus and call centers | Piloting, plans to roll out in 500 restaurants across brands |
| Popeyes U.K. | AI-powered drive-thru ('Al') | Voice ordering | Pilot reported 97% accuracy |
| Panda Express | SoundHound | AI drive-thrus | Added to ~30 locations |
| Chipotle | Conversational AI | AI-powered voice assistants for phone orders | Began testing in early 2018 |
๐ ๏ธ Technical Deep Dive
- Apprente's technology, acquired by McDonald's, was designed for complex, multilingual, multi-accent, and multi-item conversational ordering, leveraging proprietary neuroscience-based AI.
- The McDonald's-IBM AI system was characterized as a wrapper around legacy IBM Watson Natural Language Processing (NLP), where a single language model handled all aspects of order taking, including speech recognition, intent parsing, menu matching, and order confirmation.
- This architecture led to systematic failures, as it did not separate probabilistic tasks (understanding messy human speech) from deterministic business rules, struggling with accents, dialects, and background noise.
- Research suggests that integrating cross-modal approaches, such as cameras tracking lip movements alongside audio, could significantly reduce word error rates in noisy environments (e.g., from 28.8% to 12.2%).
- McDonald's new AI strategy with Google Cloud includes the deployment of AI-powered accuracy scales that weigh orders to verify contents against expected weights, and an in-house computing platform called "Edge" to support smart devices within restaurants.
- Competitor solutions often utilize generative AI technologies to process and convert customer orders, typically with a human operator available for intervention when the AI encounters difficulties.
๐ฎ Future ImplicationsAI analysis grounded in cited sources
Fast-food drive-thrus will increasingly adopt multi-modal AI solutions combining voice, vision, and weight sensors.
The past challenges with voice-only AI accuracy, coupled with McDonald's current move towards AI accuracy scales and potential for camera integration, indicate a shift towards more robust, multi-sensory AI systems for order verification and customer interaction.
The competitive landscape for AI drive-thru technology will intensify, leading to more specialized AI vendors and partnerships.
McDonald's decision to end its IBM partnership and explore other solutions, alongside the diverse range of AI providers used by competitors (Presto, Valyant AI, SoundHound, ConverseNow, Google Cloud, Nvidia), suggests a dynamic market with continuous innovation and vendor switching.
AI will shift fast-food labor roles from order-taking to order fulfillment and customer service intervention.
The primary goal of AI in drive-thrus is to automate order taking, allowing human staff to focus on preparing food, ensuring accuracy, and stepping in when AI systems encounter difficulties, thereby optimizing labor allocation.
โณ Timeline
2017
Apprente, a voice technology startup, is founded.
2019-09
McDonald's acquires Apprente, forming McD Tech Labs to develop voice-based conversational AI.
2021-10
McDonald's sells McD Tech Labs to IBM and partners to accelerate Automated Order Taking (AOT) technology.
2022-06
Reports indicate McDonald's AI voice ordering accuracy in test locations is in the low 80% range.
2024-06
McDonald's ends its partnership with IBM on AI drive-thru technology, seeking broader solutions.
2026
McDonald's plans to deploy AI-powered accuracy scales and new Google Cloud voice chatbots in drive-thrus.
๐ Sources (27)
Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
- Google Search Source
๐ฐ
Weekly AI Recap
Read this week's curated digest of top AI events โ
๐Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: The Verge โ