Throughput

Throughput is the number of tasks an AI employee can complete per unit time — the ceiling of how much work you can route through it.

Throughput is bounded by model rate limits, tool API limits, and mission complexity. In practice most AI employees can sustain hundreds of routine tasks per hour and dozens of complex ones. Hitting the ceiling is rare for a single deployment; it shows up at scale.

Example

An AI SDR can sustain ~200 outbound emails per hour, bounded by a 3-RPS Gmail send limit. Doubling the throughput requires a second sending domain.

How OpenLabor uses it

OpenLabor parallelizes within rate limits and shows you the throughput ceiling per mission.

Will the AI Employee scale infinitely?

No. External APIs and rate limits cap throughput. Plan for the ceiling at the volume you target.

Related: latency, cost-per-task.

AI Labor Glossary