Fallback Tool
A backup an agent switches to when its first-choice tool fails or isn’t available.
Think of it like
Taking the stairs when the elevator’s out — slower, but you still get up.
Example
The premium search API is down, so the agent falls back to a basic web search to keep the task moving.
How it actually works
Fallbacks add resilience by giving the agent a plan B when a tool is broken, throttled, or missing. The tradeoff is usually quality or cost — the backup is cheaper, slower, or less capable — so you accept a worse result over no result. The subtle risk is silent degradation: the agent quietly downgrades and nobody realizes the answers got worse until it matters.
For product teams
Fallbacks keep agents working through outages, but watch for silent quality drops when the backup kicks in.
For engineers
On primary-tool failure, route to an alternate tool with equivalent intent but different cost or capability.
Related
- Tool Error Handling — What triggers reaching for a fallback.
- Retry Logic — The retries that come first.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome