Enterprise tool-use training data

Use enterprise tools.Complete real workflows.

Train across connected enterprise systems. Complete complex workflows through MCP, with runnable environments and verified rewards.

Try 5 Samples
5,000tasks available
70+enterprise runtimes
8enterprise domains
Connected systems
  • records
  • documents
  • workflows
Verified result
Runnable environments. Verified rewards.

The MCP catalog

Choose your enterprise domain.

Business workflows

Customers and teams

4enterprise domains
  • Marketing + retail
  • Customer service
  • Finance
  • Human resources

Connect the evidence across business applications. Complete the work and leave the right result behind.

Operations workflows

Systems and services

4enterprise domains
  • IT service management
  • ML operations
  • Site reliability
  • Personal assistant

Work across operational systems. Resolve complex tasks with the tools and context a real deployment would use.

Choose your domains and the volume you need across 5,000 available MCP tasks.

Inside an MCP task: Vendor contracts

An empty search isn't the whole story.

Find this quarter’s vendor contract problems across six applications. Write one spreadsheet row for each contract.

Six applicationsEvidence across systems
Search + list toolsRead the text fields
Spreadsheet rowsOne row per contract
18 correct referencesComplete, accurate retrieval

Search misses evidence that the list tools expose.Task workflow

Read the task and verification summary

Find the vendor contract problems for the quarter using evidence across six applications. Search results alone can miss relevant evidence, so the agent must also use list tools and read the text fields. Write one spreadsheet row per contract.

The sample checks complete, accurate retrieval against 18 correct references. Inspect the task specification and reward verifier to understand how the final rows are evaluated.

Request the complete sample package ↗

Included with your tasks

Everything needed to run the work.

Inspect the complete task package in your samples.

Task specification
Runnable environment
Reward verifier
Customer-specific QA

Enterprise tool-use catalog evaluation

Hard tasks. Room to improve.

Frontier models pass 27–41% of the catalog’s tasks at pass@1. These results cover the broader enterprise tool-use catalog, evaluated in August 2026.

3models evaluated
pass@1success metric
Explore the full catalog ↗
41%best pass@1
Claude Fable 541%
GPT-5.6 Sol33%
Gemini 3.5 Flash27%

Your MCP mix. Your volume.

Choose your enterprise domains and the task volume you need.

1Review samples2Choose mix + volume3Get started now