# Zapier's AutomationBench grades AI agents on finished workflows across 47 apps

> Original title: What did the agent actually get done? @zapier's AutomationBench puts that question to the test across 47 simulated apps, grading the res...

- Source: Fireworks
- Published: 2026-10-28T17:14:34.000Z
- Bad Influence: https://badinfluence.news/items/mu4bw6mzfz1jgujd7fd6sso3e
- Original: https://x.com/FireworksAI_HQ/status/2115492434239230342

## Summary

Zapier's AutomationBench tests AI agents across 47 simulated apps spanning sales, marketing, operations, support, finance, and HR. The benchmark grades whether agents correctly finished each workflow, based on the results they leave behind. Fireworks AI has added AutomationBench to its Specialized Intelligence Index.

## Post

What did the agent actually get done? @zapier's  AutomationBench puts that question to the test across 47 simulated apps, grading the results agents leave behind.

Thank you to the Zapier team for partnering with us to bring this benchmark to the Specialized Intelligence Index!

Explore how they built it: https://zapier.com/blog/automationbench-on-specialized-intelligence-index/

## Quoted post @zapier

> Zapier's AutomationBench is now on @FireworksAI_HQ's Specialized Intelligence Index!
> 
> Agents work across 47 simulated apps spanning sales, marketing, ops, support, finance, & HR
> 
> We grade whether they correctly finished the workflow, so teams can pick models on what gets done 🤝

https://x.com/zapier/status/2115474627619352977