Benchmarks before promises.

See what was measured, under which conditions, and what it means for your agency. Start with a historical test or use the guides to measure your own work.

Historical benchmark

Carrier Portal Automation Benchmark

Relay scored head-to-head against leading general-purpose AI agents on real HO-3 carrier portal work.

The benchmark measures AI Quotingas it was tested in June 2026. Read the methodology and limits; this does not measure the current Workbench release.

June 2026

HO-3 carrier portal run

Field accuracy, median time to quote, and output tokens, scored against ChatGPT Codex, Claude Code, and Gemini CLI.

91.9%

field accuracy on live portal work

3.6 min

median time to quote

1st

of 4 AI agents tested head-to-head

Read the benchmark

Field notes

The work behind the numbers.

01What is a quoting AI agent? Start with the resultUnderstand quoting AI through the information it uses, actions it performs, results it returns, and the review your agency still needs.4 min 02Commercial quoting automation: prepare a file the team can useA practical commercial-lines workflow for applications, exposures, loss history, market requirements, quote status, and reviewed proposals.4 min 03Comparative rater or portal automation: compare the workflowUnderstand rating connections and browser automation by testing inputs, result status, remaining work, and recovery on your actual accounts.3 min

Your next step starts with one email.

Join the waitlist. We’re inviting agencies regularly and will email you when access is available.

Join the waitlist