Calif. Sheriff's Department to Answer Non-Emergency Calls with AI

Why It Matters

By automating routine inquiries, the agency reduces dispatcher fatigue, speeds response for critical 911 calls, and creates a scalable model for other jurisdictions facing staffing shortages.

Key Takeaways

• AI handles 20% of non‑emergency calls immediately
• Hold times dropped from 211 to 128 seconds
• $50k upfront, $150k annual AI operating cost
• Remote dispatcher team will cover half of calls
• System routes callers to correct agency, reducing misdirected calls

Pulse Analysis

The San Diego County Sheriff’s Office has become one of the first U.S. law‑enforcement agencies to embed conversational AI into its non‑emergency call center. By deploying Hyper’s natural‑language platform, the department can instantly answer callers, parse full statements, and forward the request to the appropriate unit without human intervention. This approach addresses a chronic shortage of qualified 911 dispatchers, freeing them to concentrate on life‑threatening incidents. Moreover, the AI’s multilingual capability and local colloquial training improve accuracy in a region where callers frequently reference neighborhoods rather than official jurisdictional boundaries.

The pilot, launched in March 2026, diverted roughly one‑fifth of the 400,000 annual non‑emergency calls away from the emergency center, cutting average hold time from 211 seconds to 128 seconds. The system required a modest $50,000 implementation fee and $150,000 yearly operating budget, costs absorbed by the technology fund. A planned remote‑dispatcher workforce of ten agents, paid at a lower rate than certified 911 staff, is expected to handle nearly 50 % of non‑emergency volume, further relieving on‑site personnel and improving dispatcher fatigue levels.

San Diego’s experiment signals a shift toward AI‑augmented public‑safety communications across the United States and Canada. As agencies grapple with recruitment challenges and rising call volumes, scalable bot solutions offer a cost‑effective bridge while preserving human oversight for critical incidents. Because all interactions are recorded and transcribed, the department retains full data ownership, enabling performance analytics and continuous model refinement. However, agencies must balance efficiency gains with public perception, ensuring transparent opt‑out options for callers who prefer a human voice, a factor that will shape future regulatory guidance.