Back to news
NewsJune 11, 2026· 2 min read

McDonald's AI drive-thru hits 90% accuracy in 5-location test

McDonald's ArchIQ system processed over 1 million drive-thru orders with 90% completion without staff intervention. The Google-backed tool also monitors restaurant operations—but the company's previous IBM pilot failed after customer complaints.

Our Take

McDonald's learned from the IBM disaster (remember $250 in unwanted nuggets?) and built operational oversight into ArchIQ, but five test kitchens and vendor-only metrics don't yet prove the system survives real-world volume or customer patience.

Why it matters

Drive-thru traffic has declined 5–8% monthly in 2025, pushing chains to automation as a cost lever. McDonald's success or failure here signals whether AI voice ordering can actually reduce labor without sacrificing accuracy or customer trust.

Do this week

QSR operators: request pilot participation data (error rates by daypart, staff time saved per location) before committing to cloud infrastructure or staff reductions tied to ArchIQ rollout.

Five restaurants now testing Google-backed ordering and operations monitoring

McDonald's introduced ArchIQ (nicknamed "Archy") during its Worldwide convention and is piloting the system at five undisclosed US locations. The system greets customers, processes orders in English and Spanish, handles order changes, and prompts pickup. Company-reported data shows ArchIQ has processed more than 1 million transactions with approximately 90% completed without staff escalation (per McFranchisee, a franchisee account on X).

Beyond drive-thru ordering, ArchIQ monitors restaurant operations and alerts managers to equipment failures (freezer outages) and kitchen bottlenecks. McDonald's is deploying Google Edge Cloud blades to participating restaurants ahead of broader rollout.

The system can also recognize repeat customers and offer their usual orders, though McDonald's has not disclosed how this feature works technically.

McDonald's is betting automation can reverse traffic decline—but the stakes are higher after IBM's failure

In 2025, McDonald's drive-thru traffic remained negative, declining between 5% and 8% month-over-month (per QSR Magazine's Drive-Thru Report, citing Revenue Management Solutions). The company's loyalty program and digital sales grew 20% and 19% respectively in 2024, but drive-thru foot traffic has not followed.

This is the second major AI ordering push. McDonald's ended a partnership with IBM in 2024 after customer complaints over order errors, including one case where the system allegedly added more than $250 of chicken nuggets to a single order. CEO Chris Kempczinski acknowledged in an internal memo that automation leaves fewer touchpoints with crew, which "raises the standard for hospitality when customers interact with staff."

Taco Bell and Wendy's have also announced AI-powered drive-thru systems. The success or failure of ArchIQ will shape fast-food's willingness to bet on voice automation at scale.

McDonald's has not announced a timeline for expansion beyond the five test locations or disclosed which restaurants are participating.

Operators should track these metrics before full rollout

The 90% no-escalation rate is measured by McDonald's alone and lacks independent validation. Before committing capital or staffing reductions, operators should request: error rates by daypart (breakfast vs. lunch vs. dinner), cost per transaction, staff time saved per location, and customer satisfaction scores against the pre-ArchIQ baseline at the same restaurants.

Franchisees considering adoption should also confirm: whether the Google Edge Cloud blade is a one-time or ongoing cost, how long the system takes to correct mistakes when escalated to staff, and whether McDonald's will provide indemnification if the system's errors damage customer relationships or trigger refund requests.

The IBM failure cost the company in reputation and customer trust. Evidence from five locations under controlled conditions is not enough to forecast performance in a franchisee's own market, staffing model, or customer demographic.

#Enterprise AI#Agents
Share:
Keep reading

Related stories