Role-Based Programme
RB0587

Microsoft Copilot Studio for Quality Management

AI Agent Testing, Evaluation & Quality Governance

IMAGE REQUIRED
Duration
8 Hours
Level
Basic
Delivery
Instructor-Led
Format
Workshop

Programme Objectives

  • Build foundational capability to test and evaluate Copilot Studio agents against defined business and quality requirements.
  • Design representative test scenarios covering response accuracy, groundedness, completeness, tool usage, exceptions, and conversational behaviour.
  • Validate knowledge-grounded responses, generative orchestration, tools, connectors, and multistep agent workflows.
  • Use structured agent evaluations and activity maps to identify defects, behavioural gaps, and regression risks.
  • Establish practical quality gates, traceability, responsible-AI reviews, and continuous-improvement practices before and after agent deployment.

Tools covered

Microsoft Copilot StudioTest ChatAgent EvaluationTest SetsGeneral Quality EvaluationCompare MeaningTool Use EvaluationKeyword MatchActivity MapKnowledge SourcesGenerative OrchestrationTools & ConnectorsAnalytics

Who should attend

  • Quality Assurance Managers
  • Quality Managers
  • QA Engineers
  • Quality Engineers
  • Software Quality Professionals
  • Test Managers
  • Test Analysts
  • UAT Professionals
  • Process Quality Professionals
  • AI Quality Analysts
  • Release Quality Professionals
  • Quality Management Team Leads

Prerequisites & Participant Readiness

  • Basic understanding of quality assurance, testing, or validation practices
  • Familiarity with requirements, expected results, defects, and acceptance criteria
  • Basic awareness of AI agents or business-process automation is helpful
  • No programming expertise required
  • No previous Microsoft Copilot Studio experience required
  • Access to specific evaluation capabilities may depend on organisational environment, licensing, and preview availability

TOC Modules

Concepts
  • Understanding agents, instructions, knowledge, topics, tools, and generative orchestration
  • Mapping quality across Build → Test → Evaluate → Publish → Monitor → Improve
  • Understanding deterministic software testing versus probabilistic AI-agent evaluation
  • Identifying agent-quality risks such as inaccurate answers, poor grounding, incorrect tool use, and unexpected routing
Practical activities
  • Exploring a sample Copilot Studio agent
  • Reviewing its instructions, knowledge, tools, and intended behaviour
  • Creating an initial agent-quality checklist

Scenarios

Enterprise Knowledge Agent to Quality Sign-Off

Business Requirements → Knowledge Sources → Test Scenarios → Test Chat → Groundedness & Quality Evaluation → Defect Analysis → Correction → Regression Test → QA Sign-Off

Participants validate an enterprise knowledge agent against approved information and defined quality criteria before recommending production release.

Action-Oriented Agent to Controlled Release

User Request → Generative Orchestration → Tool / Connector → Business Action → Exception Tests → Tool-Use Evaluation → Regression Validation → Quality Gate

Participants evaluate an agent that performs business actions, verifying correct tool selection, inputs, failures, confirmations, and escalation behaviour before controlled deployment.

Related programmes

Continue with programmes from the same capability area.

Take the next step

Ready to make this programme work for your team?

Customise modules, duration and business scenarios for your team.

Instructor-ledVirtualHybrid

Designed around your roles, tools and real workflows.