Table Side AI

 

Artificial Intelligence, Generative Research, evaluative Research, interaction Design, visual design, Prototyping

 

SITUATION

Darden Restaurants, a leader in the dining industry with brands ranging from casual to fine dining, sought to improve server efficiency and order accuracy while maintaining a high-quality customer experience. By identifying pain points in server workflows, we aimed to reduce wait times and enhance overall service quality.

HYPOTHESIS

If server efficiency and accuracy are improved by reducing reliance on the Point of Sale (POS) system, then servers will be able to spend more time engaging with guests, reduce wait times, and increase customer satisfaction and tips.

Research Methodology

A generative research approach was undertaken to understand server pain points and identify opportunities for improvement. This involved:

  • Field Observations: Shadowing 2-3 servers per location with an even tenure distribution during lunch and dinner shifts.

  • Ad-hoc Inquiry: Engaging servers in discussions during back-of-house moments to understand their frustrations and strategies.

  • Research Scope: 48 servers across Darden brands, spanning casual to fine dining establishments in the Southwest, Midwest, and Southeast.

Key Findings

  1. POS System Bottlenecks

    • The POS system was frequently unavailable, down, or difficult to use.

    • Multiple trips to the POS slowed down service, increasing the time from order placement to food delivery.

  2. Impact on Service Quality

    • Servers struggled to maintain rapport with guests due to frequent POS interactions.

    • In fine dining, technology interruptions diminished the elevated guest experience.

    • Servers expressed concerns about carrying a clunky device that further slowed them down.

  3. Efficiency Concerns

    • POS inefficiencies had a far greater negative impact than initially speculated.

    • Reducing the number of POS interactions could significantly improve workflow efficiency.

Solution: Voice-Enabled Ordering System

To address these issues, we leveraged Google’s Speaker Diarization AI model, enabling servers to send orders to the kitchen in real-time and:

  1. Identify and differentiate between multiple speakers, ensuring order accuracy.

  2. Move seamlessly between guests by voice, eliminating unnecessary POS interactions.

  3. Reduce order placement time, cutting the standard wait time by approximately 10 minutes.

Results & Impact

Implementing a voice-enabled ordering system proved to be a transformative step in enhancing server efficiency and guest experience.

  • Servers were able to focus more on guest interactions, improving rapport and overall service quality.

  • The reduced reliance on physical POS interactions streamlined workflows, significantly cutting down service bottlenecks.

  • Fine dining experiences remained elevated, as technology seamlessly integrated into the service process rather than disrupting it.