DoorDash Scales Customer-Centric AI Across 19,000 Agents With Nearly 100% Automated Quality Coverage

As a global leader in local commerce, DoorDash operates at an extraordinary scale. Millions of customer conversations take place across its ecosystem every month, spanning consumers, merchants, and Dashers. With approximately 19,000 support agents operating across internal teams and BPO partners, maintaining consistent, high-quality interactions is both mission-critical and operationally complex.

Historically, quality assurance in contact centers relied on manual review of a small percentage of interactions. For a company moving at DoorDash’s pace, that model created blind spots.

DoorDash set out to fundamentally change that.

Through a strategic partnership with Observe.AI, powered in part by AWS AI infrastructure, DoorDash automated interaction evaluations across nearly 100% of customer conversations. What was once limited to small sampling became comprehensive, consistent, and scalable.

“We knew manual QA wasn’t going to scale with our growth,” said Xenia Strunnikova, Head of Customer Experience, Fraud, Trust & Safety S&O at DoorDash. “If we truly wanted to understand our customer experience, we needed visibility into every interaction—not just a small subset. Achieving near 100% coverage was a game changer.”

Automation did not replace human oversight—it amplified it. By removing the burden of manual scoring, DoorDash enabled its quality and operations teams to focus on higher-value analysis, coaching, and customer protection.

The result: enterprise-wide visibility, consistent standards across partners, and a scalable foundation for long-term AI innovation.

Moving Beyond Checkbox QA to Understand Behavioral Drivers

For DoorDash, automation was not just about efficiency. It was about evolution.

Traditional QA often centers on binary, compliance-driven scorecards—did the agent say the right phrase, follow the right script, check the right box. But customer experience is rarely binary. It is nuanced, emotional, and behavioral.

DoorDash wanted to understand why customers felt the way they did—not just whether agents followed protocol.

By leveraging Observe.AI’s conversational intelligence signals—including sentiment analysis, comprehension metrics, and behavioral indicators—DoorDash shifted from surface-level compliance monitoring to deeper diagnostic insight.

|“We made a conscious decision to move away from checkbox QA,” said Joaquin Dufeu, Director of Strategy & Operations for Customer Experience & Integrity (CXI) with DoorDash. “Our goal was to better understand the drivers behind customer sentiment. What’s creating frustration? Where are we missing empathy? Where are policies creating friction? AI allowed us to see those patterns at scale.”

The platform enables DoorDash to infer customer satisfaction even when direct signals—like survey responses—are missing. By analyzing tone, language, and interaction dynamics, teams can surface subjective issues that previously went undetected.

|This shift transformed QA from a monitoring function into a strategic lever for improving customer experience.
‍

Accelerating Hotspot Detection From Weeks to Near Real Time

One of the most powerful outcomes of full coverage and advanced signal detection has been speed.

In a dynamic marketplace like DoorDash’s, new product features, policy changes, or subscription programs such as DashPass can quickly generate customer confusion or friction. Previously, identifying those issues required days or weeks of investigation—often triggered only after survey trends or escalations signaled a problem.

Now, conversational intelligence surfaces emerging patterns almost immediately.

“Listening to customers at scale is incredibly difficult,” said a third DoorDash leader. “With automated signals and near real-time visibility, we can identify hotspots we might never have seen before. What used to take days—or sometimes weeks—can now surface almost immediately.”

This acceleration enables proactive intervention. Teams can quickly align cross-functionally, flag product friction points, and refine policies before issues compound.

The organization is now moving toward a future-state vision of real-time operational response—where insight drives action in the moment, not after the fact.

For a company investing heavily in AI across its broader business, this capability reinforces DoorDash’s ambition to lead responsibly in AI-driven customer experience.

Strengthening BPO Partnerships With Objective, Data-Backed Insight

DoorDash’s customer support model includes multiple global BPO partners. As the organization scaled, ensuring consistency and alignment across partners became increasingly important.

Manual QA models often create tension—limited samples can lead to debates over scoring, fairness, and performance interpretation. Automated, comprehensive evaluation changes that dynamic.

With a shared, objective view of interaction data, DoorDash and its partners are now aligned around evidence-based insight rather than anecdotal samples.

“Instead of debating individual calls, we’re aligned around patterns and trends,” said Joaquin Dufeu, Director of Strategy & Operations for Customer Experience & Integrity (CXI) with DoorDash. “That transparency strengthens our partnerships. It creates fairness, accountability, and a shared commitment to improvement.”

By leveraging AWS AI infrastructure for secure, scalable transcription and model support, and Observe.AI’s intelligence layer for analysis and signal detection, DoorDash created a consistent framework for evaluating quality across its global footprint.

“DoorDash’s approach demonstrates how AI can strengthen—not strain—enterprise ecosystems,” said an AWS executive. “By combining AWS AI infrastructure with Observe.AI’s conversational intelligence platform, DoorDash is scaling securely while maintaining high standards for customer experience and operational integrity.”

The result is improved coaching consistency, clearer performance expectations, and reduced operational friction across partners.

Empowering Teams to Focus on Customers, Not Manual Scoring

Automation reshaped the role of DoorDash’s quality and operations teams.

Instead of spending hours manually reviewing calls and assigning scores, teams now use AI-generated insights as a starting point for coaching, calibration, and deeper analysis.

“We see AI as an amplifier,” Xenia Strunnikova, Head of Customer Experience, Fraud, Trust & Safety S&O at DoorDash explained. “It gives us better signals so our teams can focus on developing people and protecting customers, rather than manually grading interactions.”

The shift also allowed DoorDash to address more subjective issues—such as empathy, tone, and customer safety—that are traditionally difficult to measure consistently.

With nearly full coverage and behavioral signals embedded into evaluation workflows, leaders gain clarity on where coaching is most needed and where systemic issues may be impacting experience.

This creates a virtuous cycle: better insight leads to better coaching, which leads to better interactions, which leads to stronger customer trust.

Advancing Responsible AI Leadership in Customer Experience

For DoorDash, this initiative is part of a larger AI roadmap.

Across the organization, AI is being explored and implemented thoughtfully to drive operational excellence and customer satisfaction. Customer support is one of the most visible and impactful applications.

Importantly, DoorDash does not view AI as a replacement for human judgment. Instead, it is an enabling platform—built iteratively in partnership with Observe.AI and AWS—to deliver intelligence at scale while keeping the customer at the center.

“This work is about investing in our customers,” said Amar Dhaliwal, Sr. Manager of Strategy & Operations at DoorDash. “We want to be on the forefront of AI in customer experience—but responsibly. That means using AI to better understand our customers, keep them safe, and make sure we’re always doing what’s right.”

The collaboration reflects a zero-to-one journey—building custom signals, refining models, and iterating together to align technology with operational realities.

Rather than deploying a static solution, the partnership continues to evolve as new features launch, new data becomes available, and new insights emerge.

Building the Foundation for Real-Time Customer Intelligence

Today, DoorDash operates with nearly 100% automated quality coverage across 19,000 agents, near real-time hotspot detection, and deeper behavioral insight than ever before.

But the journey is ongoing.

The companies are continuing to refine sentiment models, expand signal sophistication, and explore ways to drive even more proactive, real-time response capabilities.

At its core, the transformation is about one thing: making it easier for DoorDash to listen—and respond—to customers at scale.

From comprehensive coverage to faster insight, stronger partnerships, and more empowered teams, the business outcomes extend well beyond QA metrics.

They represent a fundamental shift in how customer experience is understood, measured, and improved.

And as DoorDash continues investing in AI across its business, the customer support transformation stands as a clear example of what responsible, customer-centric AI can achieve when built through deep partnership.

‍