One Brand's Chatbot Drives 75% of All Sales. The Other Scrapped a 3-Year Project.
Domino's and McDonald's are the two most recognized restaurant brands on Earth. Both invested heavily in AI chatbots. Both had bold visions for how automation would transform ordering. But their paths diverged in ways that offer powerful lessons for every restaurant, regardless of size.
Domino's launched "Dom," a multi-platform chatbot that now supports ordering across Facebook Messenger, WhatsApp, SMS, Amazon Alexa, and Google Assistant. Today, over 75% of Domino's sales come through digital channels, and the chatbot has delivered a 70% reduction in customer query resolution time.
Meanwhile, McDonald's partnered with IBM in 2021 to deploy AI-powered voice ordering across more than 100 U.S. drive-thru locations. By July 2024, they pulled the plug entirely after persistent failures with accents, background noise, and complex modifiers. The company then pivoted to a Google Cloud partnership.
- 75%+
Of Domino's sales are digital
- 70%
Reduction in query resolution time
- 100+
McDonald's AI locations before shutdown
- 3 yrs
Years McDonald's spent before scrapping AI
The Domino's Playbook
Domino's Dom: How a Pizza Chain Became a Tech Company
Domino's didn't build Dom overnight. The company has positioned itself as a technology company that happens to sell pizza, investing in digital infrastructure for over a decade. Dom is the customer-facing result of that strategy: a chatbot that meets customers wherever they already are.
Domino's Digital Evolution
2015
Dom launches on Facebook Messenger
First major pizza chain to offer full ordering through a social messaging platform
2016-2018
Multi-platform expansion
Added Amazon Alexa, Google Assistant, SMS, Twitter, and smart TV ordering
2019-2021
AI-powered conversations
Upgraded from scripted flows to natural language processing for more flexible ordering
2022-2024
WhatsApp integration
Added WhatsApp ordering to reach international markets where the app dominates
2025-2026
75%+ digital sales achieved
Digital channels now drive the majority of all Domino's revenue worldwide
What makes Dom successful isn't the technology itself. It's the strategy behind it. Domino's understood three things early: customers don't want to download another app, ordering should happen where people already spend time, and reducing friction by even one step dramatically increases conversion rates.
Meet Customers Where They Are
Dom is available on 6+ platforms. Instead of forcing customers to download a proprietary app, Domino's went to Messenger, WhatsApp, Alexa, and SMS.
Start Simple, Iterate Fast
Dom started with basic scripted ordering on Messenger and evolved to NLP-powered conversations over several years. No big-bang launch.
Track Everything, Optimize Constantly
Domino's obsessively tracks digital order metrics. The 70% reduction in query resolution time came from continuous optimization, not a single deployment.
Reorder Is the Killer Feature
Dom remembers previous orders and enables one-tap reordering. This single feature drives massive repeat purchase rates across all channels.
The McDonald's Cautionary Tale
McDonald's Automated Order Taker: A $200 Billion Company's Most Public AI Failure
In 2021, McDonald's launched its Automated Order Taker (AOT) in partnership with IBM, deploying AI-powered voice ordering across more than 100 drive-thru locations in the United States. The system was designed to take spoken orders, handle modifications, and upsell items, all without human intervention.
The failures were spectacular and very public. Viral videos showed the system adding nine sweet teas to a single order, suggesting bacon on ice cream, and struggling with basic modifications like "no pickles." Customers with regional accents or who ordered in noisy environments found the system nearly unusable. By July 2024, McDonald's ended the IBM partnership entirely.
What Went Wrong at McDonald's
Voice AI in noisy environments
Challenge #1
Drive-thrus have wind, traffic, music, and multiple passengers speaking simultaneously
Regional accents and languages
Challenge #2
The U.S. market has enormous linguistic diversity the system couldn't handle
Complex modifications
Challenge #3
Real orders have substitutions, removals, additions, and special requests the AI misinterpreted
Public viral failures
PR Crisis
Social media amplified every mistake, eroding customer trust faster than fixes could ship
IBM partnership dissolved
Project Killed
After 3 years of investment, the entire project was scrapped in July 2024
Key insight Even $200B+ companies fail at custom AI
Critically, McDonald's didn't give up on AI. The company pivoted to a Google Cloud partnership and launched "Ask Pickles," an internal AI assistant for crew members. The lesson wasn't that AI doesn't work. It was that the approach mattered: customer-facing voice AI in noisy environments was too ambitious, while text-based ordering and internal AI tools proved far more reliable.
More Brand Lessons
What Other Major Chains Are Doing Right (and Wrong)
The chatbot landscape across major restaurant chains reveals a clear pattern: text-based bots on existing messaging platforms consistently outperform custom-built voice AI. Here's what several other brands have learned.
Big Brand Chatbot Strategies at a Glance
| Brand | Approach | Result |
|---|---|---|
| Domino's (Dom) | Multi-platform text bot (Messenger, WhatsApp, SMS, Alexa) | 75%+ digital sales, 70% faster query resolution |
| McDonald's (AOT) | Custom voice AI at drive-thrus with IBM | Scrapped after 3 years; pivoted to Google Cloud |
| Pizza Hut (Hutty) | Messenger and Twitter chatbot for ordering | 2x purchase completion rate vs. web, 2x faster orders |
| Wendy's (FreshAI) | Google Cloud voice AI partnership at drive-thrus | Expanded to select locations; still in testing phases |
| White Castle | SoundHound voice AI at drive-thrus | Rolled out to 100+ locations with reported improvements |
| Starbucks | AI for personalization and back-of-house operations | Focus on internal efficiency rather than customer-facing ordering |
5 Lessons Every Restaurant Can Apply Today
You don't need a billion-dollar budget to apply the strategies that work for the biggest brands. The lessons from their successes and failures distill into five actionable principles for restaurants of any size.
What the Big Brands Teach Us
Contrasting the strategies that worked vs. those that failed
What Failed
Starting with voice AI
McDonald's proved that voice ordering in uncontrolled environments is unreliable, even with IBM-level resources.
Building everything custom
Multi-year, multi-million-dollar custom builds that were outdated before they launched.
Big-bang launches
Deploying AI across 100+ locations before proving it works consistently at even one.
What Worked
Starting with text-based bots
Domino's and Pizza Hut proved that messaging platforms deliver higher conversion with lower complexity.
Using established platforms
Deploying on WhatsApp, Messenger, and SMS where customers already are, not forcing new app downloads.
Iterative rollouts
Starting with one channel, measuring results, then expanding based on real data.
The 5 Principles from Big Brand AI
AVOID THIS
- Don't start with voice AI in noisy environments
- Don't build custom from scratch unless you have $500K+ budget
- Don't launch across all locations before proving the concept
- Don't ignore customer feedback and viral complaints
- Don't choose technology before choosing the right channel
DO THIS
- Start with text-based chatbots on messaging platforms
- Deploy where customers already spend time (WhatsApp, Messenger, SMS)
- Enable easy reordering (Domino's killer feature)
- Iterate based on data, not assumptions
- Use third-party platforms to launch fast, then customize
How to Apply These Lessons Without the Enterprise Budget
The irony of these case studies is that independent and small-chain restaurants can actually implement the winning strategies faster than the brands that pioneered them. Domino's spent a decade building its multi-channel presence. Today, an AI-powered ordering agent can deploy across WhatsApp, Messenger, and Instagram in under two weeks through platforms that didn't exist when Dom first launched.
| Current monthly orders | 1,000 orders |
|---|---|
| Average order value | $35 |
| Digital conversion lift (industry average) | 25% |
| Additional monthly orders from AI channel | 250 orders |
| Potential additional monthly revenue | $8,750 |
Based on industry-average 25% increase in completed orders during peak times with AI chatbot ordering | |
Your Turn
Get Domino's-Level Multi-Channel Ordering for Your Restaurant
Ordering.co deploys AI ordering agents across WhatsApp, Messenger, and Instagram in under two weeks. No custom development, no multi-year projects, no IBM partnership needed.
Frequently Asked Questions
Common questions about restaurant chatbot strategies
What chatbot does Domino's use for ordering?
Domino's uses "Dom," a multi-platform chatbot available on Facebook Messenger, WhatsApp, SMS, Amazon Alexa, and Google Assistant. Dom handles full ordering flows including menu browsing, customization, reordering, and order tracking. Over 75% of Domino's sales now come through digital channels.
Why did McDonald's scrap its AI ordering system?
McDonald's ended its IBM-powered Automated Order Taker (AOT) in July 2024 after deploying it at 100+ U.S. drive-thrus. The voice AI struggled with regional accents, background noise, and complex order modifications. Viral videos of ordering errors (nine sweet teas, bacon ice cream) damaged customer trust. McDonald's pivoted to a Google Cloud partnership focused on different AI applications.
Is voice AI or text-based chatbot better for restaurants?
Text-based chatbots on messaging platforms consistently outperform voice AI for restaurants. Domino's and Pizza Hut have proven this with 75%+ digital sales and 2x purchase completion rates respectively. Voice AI in noisy environments (drive-thrus, kitchens) remains unreliable, as McDonald's demonstrated. Start with text, and consider voice only after text channels are optimized.
How did Pizza Hut's chatbot perform compared to web ordering?
Pizza Hut's Messenger chatbot "Hutty" showed that Facebook users were 2x more likely to complete purchases compared to web orders, and orders processed 2x faster through the bot than traditional channels. The frictionless messaging experience removed barriers that web-based ordering creates.
Can a small restaurant replicate what Domino's does with chatbots?
Yes. The core Domino's strategy (multi-channel presence, frictionless ordering, easy reordering) is now accessible through third-party AI platforms at a fraction of the cost. What took Domino's a decade to build, platforms like Ordering.co can deploy in 1-2 weeks across WhatsApp, Messenger, and Instagram.
What is McDonald's doing with AI after scrapping the IBM project?
McDonald's pivoted to a Google Cloud partnership and launched "Ask Pickles," an internal AI assistant that helps crew members with operational tasks. Rather than customer-facing voice ordering, McDonald's shifted to using AI for back-of-house efficiency, menu optimization, and crew support, a more pragmatic approach to AI adoption.
Which restaurant chains have successful chatbot programs?
Domino's (Dom), Pizza Hut (Hutty), and White Castle (SoundHound) have the most successful documented programs. Wendy's (Google Cloud FreshAI) and Burger King (Patty) are in testing phases. Starbucks uses AI primarily for personalization and back-of-house operations rather than direct ordering.
What is the biggest lesson from big brand chatbot deployments?
Start with text-based chatbots on platforms where your customers already are. Every successful deployment (Domino's, Pizza Hut) started with messaging platforms, not custom-built voice AI. The failures (McDonald's AOT) came from attempting the hardest problem first. Text-based bots are proven, lower-risk, and deliver measurable results faster.