When orders went cold on the line
It was dark outside when Nina finished closing the coffee shop. She was tired, starving, and not in the mood to cook. She ordered her usual burrito from a local spot that wasn’t on the DoorDash app. The system placed a call to the restaurant to confirm it. But when someone picked up, the kitchen noise made it hard to hear. The line cut out. Her dinner never came.
DoorDash runs a massive, always-moving ecosystem with over 500,000 merchants, five million Dashers, and tens of millions of customers. But in 2019, too many phone-based food orders were being placed inefficiently by offshore call centers. This led to missed revenue and rising operational costs. Cancellations and escalations added to the load. In 2023, a new challenge emerged on the Dasher side. A spike in tax-related calls overwhelmed systems every January.
In both cases, DoorDash needed a way to scale instantly without hiring seasonal agents or expensive BPOs. So they turned to Replicant.
"Our goal was to find a high quality solution to handle the erratic demand of our phone order traffic, while minimizing escalations and continuing to lower the costs," said George McConnell, Former Global Operations Manager at DoorDash.
Clear calls, hot meals
DoorDash first partnered with Replicant to automate outbound food ordering. This was a tedious, high-volume task that was essential to customer satisfaction but difficult to scale. Replicant’s AI agents were designed to clearly and succinctly communicate order details, even through background noise, unique accents, or voicemail trees. The goal was to match or outperform human agents in resolution, speed, and clarity.
Replicant’s AI agents also tested and optimized how and when to retry calls that went to voicemail or came back busy. This dramatically reduced order failure rates and brought formerly disengaged merchants back into the flow.
In 2023, DoorDash expanded the partnership again. This time they automated a massive volume of outbound tax form verification calls. The campaign launched in just two weeks and scaled to over 2,000 concurrent calls at peak. Dashers got the support they needed faster. DoorDash avoided hiring hundreds of seasonal agents during the holiday crunch.
"Replicant was able to meet or exceed our goals across all major metrics but more importantly, our partner restaurants preferred the AI solution," said McConnell. "During our initial pilot we spoke with numerous merchants who explained they preferred speaking with the AI agent since the kitchen is often noisy and the robot is clear."Â
Faster service, full satisfaction
In just six weeks, Replicant’s AI agents were placing over 35,000 food orders per day. Resolution rates climbed. Escalations dropped. Customer satisfaction went up. Operational efficiency did too.
For the tax season campaign, Replicant handled nearly 800,000 calls in just eight days and saved more than 2.5 million agent minutes. That kind of scale would have been impossible with live support alone.
Now when someone like Nina places an order after a long shift, it goes through. Even if the kitchen is loud, the restaurant gets the call, the order is placed clearly, and the food shows up hot. Fewer missed meals. Fewer dropped calls. More satisfied customers at the end of their day.
DoorDash now uses Replicant’s AI agents as a flexible, always-on extension of their team. It delivers cost savings and faster service without sacrificing experience.
Transformational results:
Outbound food ordering
- 350,000+ calls automated per day
- 87 percent resolution rate
- 6 weeks from kick-off to launch
- 0-minute hold times
