93% Accuracy Achieved in Simulating Food Waste Reduction Policies

Highlights

• Achieved a 40% reduction in food waste in simulated environments, closely mirroring the real-world reduction of 43%.
• Empowered stakeholders to test and refine policy interventions before real-world implementation, minimizing costs and risks.
• Validated the use of virtual pilots to scale effective interventions in different areas.

Case Study

Using Decisions Lab's AI simulation tool, we've replicated and evaluated the effectiveness of a food waste reduction policy implemented in Taikoo Shing restaurants. The policy focused on leveraging social marketing nudges to influence diners' decisions and reduce waste. Key measures included offering diners options to adjust portion sizes, opt out of unwanted garnishes, and engage with informational prompts. The original intervention achieved a 43% reduction in weekday food waste, and our simulation aimed to replicate and analyze these results in a controlled virtual environment.

Simulation Design

• Stakeholders Modeled: Simulated behaviors of diners, including their decision-making processes and interactions with restaurant workflows.
• Core Actions Tracked: Ordering patterns, responses to nudges, and resulting waste generation.
• Evaluation Metrics: Daily food waste reduction rates, behavior shifts, and simulated survey feedback.

Results

The simulation achieved a 40% reduction in food waste, demonstrating a 93% alignment with real-world outcomes. This close correspondence underscores the potential of Decisions Lab’s LLM-powered simulation technology to accurately evaluate and optimize policy interventions before deployment. By providing actionable insights and scalable solutions, virtual tools like ours enable stakeholders to reduce risks, maximize impact, and accelerate sustainable change.