How this is calculated:
Each Ninth District question is estimated to represent approximately 5 minutes of staff time, such as answering a phone call, responding to an email, or assisting a walk-in question.

Total Ninth District questions × 5 minutes ÷ 60 = estimated hours saved.

This is a conservative estimate. Some questions may take longer, and after-hours questions may represent issues staff would otherwise need to address the next business day.

Estimated Anthropic API cost for this period, based on actual input/output tokens logged per question.

Calculated as: (input tokens × $3 + output tokens × $15) ÷ 1,000,000.

This reflects model usage cost only. With prompt caching active, cached input is billed at a reduced rate, so real cost may be lower than shown.