Summary
DoorDash (NASDAQ: DASH) improved its support workflows with Giga's AI agents in just weeks, now maintaining over 90% DWR (Did We Resolve) rate, based on data collected from September 20, 2025 to October 20, 2025 while handling complex edge cases.
Contents
Share
While enterprise AI deployments can often span quarters, Giga went from integration to production-scale impact in weeks.
The accelerated timeline included rapid integration with DoorDash's core infrastructure without disrupting live operations. Giga's AI agents trained on historical edge cases while shadowing live support interactions, calibrating reasoning against DoorDash's operational policies.
Within weeks, the system progressed from pilot to full production, expertly handling complex multi-party orchestrations.
At DoorDash, we operate at a massive scale across services, platforms, and languages. Giga leveraged usage data to deliver measurable improvements, including fewer escalations, faster resolution paths, and more efficient workflows across our teams. As we continue to grow across more than 40 countries and serve nearly 50 million people each month*, partnerships like this are critical to delivering better outcomes for consumers on a global scale.
Andy Fang
Co-Founder at DoorDash
* Based on the number of individual consumer accounts that have completed an order on the DoorDash, Wolt and Deliveroo marketplaces during December 2024.
DoorDash is one of the most operationally complex businesses in the world:
More than 10 billion lifetime orders
Over 500,000 total merchants on the DoorDash Marketplace
In 2024, more than 8 million Dashers used the platform
Hundreds of thousands of requests for assistance per day
Even small improvements in response time add up. Reducing a few minutes per support interaction translates into faster deliveries and a smoother experience across DoorDash’s network.
DoorDash's operational complexity extends far beyond volume. Each delivery requires real-time coordination across multiple systems: merchant timing, Dasher routing, consumer availability, payment processing, and fraud detection — all while moving quickly to help ensure the best order quality, particularly for food deliveries.
The geofence mismatch exemplifies this challenge. When a Dasher can't complete delivery at the designated address, the AI must become a real-time investigator:
Discovery: Diagnose why the Dasher is blocked from completion
Analysis: Determine if the consumer requested a new address outside the delivery zone
Investigation: Search chat history between Dasher and consumer for address change requests
Verification: Call the consumer if no documentation exists
Confirmation: Verify successful delivery to the new location
Resolution: Apply policy-compliant decision to mark the order complete
Previously, this single edge case could trap Dashers on hold and delay subsequent deliveries.
Multiply this across thousands of concurrent problems — wrong addresses, closed restaurants, payment failures — and support can become a bottleneck that compounds delays throughout the entire network.
Giga's AI agents address these challenges through parallel orchestration, managing both sides of an interaction simultaneously:
Maintaining live connection with the Dasher to capture real-time context
Calling the consumer to verify the updated address
Cross-checking both inputs against DoorDash's operational policies
Providing guidance to the Dasher to help complete or reroute within compliance
What once may have been a time-intensive task has become a real-time reasoning loop that helps uphold compliance, transparency, and speed — improving DWR.
Giga’s AI agents now support DoorDash’s live-delivery scenarios at production scale with over a 90% DWR rate, based on data collected from September 20, 2025 to October 20, 2025.
Together with DoorDash, we're tackling the most complex support workflows in the world and targeting 98% resolution rate while advancing the next generation of enterprise support.
GET A PERSONALIZED DEMO
Giga’s AI agents handle complex workflows at scale, from live delivery issues to compliance decisions, while maintaining over 90% resolution accuracy in production.
