Scope paragraph
This result applies to the support-ticket draft workflow tested on the recorded date, model version, prompt version, tool scope, and scenario set.
Resource
A sample appendix for explaining a completed assessment to a customer: the workflow tested, the configuration, the methods, the results and the limits. Swap the example text for your real report details before you share it. And to be clear, no AgentBoundary certifications have been issued or published.
Appendix contents
Sample paragraphs
Fill in the workflow, configuration, test date, results and limitations from the real report, and say which changes would call for another review. The sample is a format. It isn't a result.
This result applies to the support-ticket draft workflow tested on the recorded date, model version, prompt version, tool scope, and scenario set.
This result does not cover future model swaps, prompt rewrites, memory behavior changes, new tools, policy changes, or major retrieval changes.
Findings remediated after testing should be documented with fix date, owner, validation method, and whether retesting is required.
01
02
The public AgentBoundary benchmark evaluates model responses through chat-completion APIs. If your assessment also tested application permissions or tool execution, those are separate methods with separate observations, and the appendix should describe them that way. A benchmark score is not a test of your deployed integration.
Related services and resources
Next step