{"title":"Test the Copilot capability in AL","url":"https://jonas-paprotka.de/docs/dev-itpro/developer/ai-test-copilot/","by":"JOP-AL","license":"Copyright (c) Microsoft Corporation. CC BY 4.0, https://creativecommons.org/licenses/by/4.0/; code samples MIT. From https://learn.microsoft.com/en-us/dynamics365/business-central/dev-itpro/developer/ai-test-copilot, MicrosoftDocs/dynamics365smb-devitpro-pb at 828b6ee (2026-11-19). Changed: includes filled in, links to this site, and for agents shortened without dropping a fact.","credit":"Microsoft Corporation, CC BY 4.0","origin":"https://learn.microsoft.com/en-us/dynamics365/business-central/dev-itpro/developer/ai-test-copilot","source":{"name":"Business Central developer and IT pro docs","repo":"https://github.com/MicrosoftDocs/dynamics365smb-devitpro-pb","commit":"828b6ee72fcd75755627cd6011f3d20899217ab2","date":"2026-11-19"},"checked":"2026-11-19","description":"Explore testing strategies for Copilot features in AL. Ensure AI-generated outputs are accurate, safe, and user-friendly.","date":"2026-10-02","meta":{"topic":"concept-article"},"outline":[{"level":1,"title":"Test the Copilot capability in AL","slug":"test-the-copilot-capability-in-al","at":0,"chars":1798},{"level":2,"title":"Why test Copilot features?","slug":"why-test-copilot-features","at":400,"chars":331},{"level":2,"title":"Permission set prerequisite for Evaluation","slug":"permission-set-prerequisite-for-evaluation","at":731,"chars":225},{"level":2,"title":"Overview of Evaluation","slug":"overview-of-evaluation","at":956,"chars":635},{"level":2,"title":"Related information","slug":"related-information","at":1591,"chars":207}],"text":"# Test the Copilot capability in AL\nCopilot features can dramatically improve user productivity, but only when they're reliable. Since Large Language Models (LLMs) generate output based on probabilistic patterns, traditional rule-based testing approaches aren't sufficient. The next articles help developers and testers adapt their strategies to ensure high-quality, trustworthy Copilot experiences.\n## Why test Copilot features?\nTesting helps ensure that Copilot features:\n Deliver accurate, relevant responses\n Respond consistently to similar user prompts\n Avoid producing harmful, biased, or inappropriate content\nWithout proper validation, AI-generated outputs can lead to user frustration, compliance risks, and brand damage.\n## Permission set prerequisite for Evaluation\nTo use Evaluation, users must have the AI TEST TOOLKIT permission set assigned. This permission set provides access to the necessary pages and actions to create and run AI tests.\n## Overview of Evaluation\nIn the next articles, we explore how to test Copilot features in Business Central using Evaluation (formerly the AI Test Toolkit). Evaluation allows you to create and run tests that validate the accuracy, safety, and reliability of your Copilot features. It also supports suite-level configuration, multilingual test execution, Copilot credit visibility, and Business Central agent evaluation scenarios.\n Create datasets<ai-test-copilot-datasets>\n Write AI tests<ai-test-copilot-ai-tests>\n Evaluation<ai-test-copilot-testtool>\n Best practices for testing the Copilot capability<ai-test-copilot-bestpractices>\n## Related information\nBusiness Central Copilot Evaluation<https://github.com/microsoft/BCApps/blob/main/src/Tools/AI%20Test%20Toolkit/README.md>\nBuild the Copilot capability in AL<ai-build-capability-in-al>"}