A is correct: The 'Ask a question' feature relies on a QnA Maker knowledge base to provide relevant AI-driven answers to agents during interactions.

B is correct: Proper role-based access control ensures that only authorized agents can use the feature, maintaining security and compliance.

C is incorrect: While sentiment analysis is a useful AI feature, it is not a prerequisite for configuring the 'Ask a question' functionality.

D is correct: Tailoring the AI model to the domain-specific language improves the accuracy and relevance of AI suggestions in the 'Ask a question' feature.

E is incorrect: Although call transcription supports other AI features, it is not directly required to configure the 'Ask a question' feature.

F is incorrect: Fallback intents are relevant to virtual agents or bots, not directly to the 'Ask a question' feature used by human agents.

A is correct: AI-generated skill tags help dynamically classify agent expertise, enabling precise routing decisions based on real-time AI assessments.

B is correct: Incorporating AI skill match thresholds allows the system to balance AI predictions with predefined skill requirements, optimizing routing accuracy.

C is incorrect: Manual skill assignments remain important to cover skills not yet captured or accurately identified by AI models.

D is correct: Real-time content analysis by AI models enables dynamic skill inference and improves routing precision.

E is incorrect: Agent availability must be accurately reflected; setting agents offline would prevent work assignment regardless of AI routing.

F is correct: Continuous learning ensures AI models evolve and improve routing decisions by incorporating real agent performance data.

A is incorrect: A standalone AI bot without orchestration cannot reliably manage call routing or context transfer to human agents, leading to poor customer experience.

B is correct: Classic orchestration acts as the central controller, coordinating generative AI capabilities and routing calls to human agents with preserved context, ensuring seamless interaction.

C is incorrect: Limiting classic orchestration to post-call functions misses its critical role in real-time call routing and context management during interactions.

D is incorrect: Operating generative AI independently without integration into orchestration disrupts context continuity and handoff efficiency between AI and agents.

A is correct: Real-time feedback allows agents to provide insights while the interaction context is fresh, improving the accuracy and relevance of feedback.

B is correct: Post-interaction surveys help gather comprehensive feedback without interrupting the agent's workflow during active conversations.

C is incorrect: Generic questions may not capture channel-specific nuances, reducing the usefulness of feedback for channel optimization.

D is correct: Leveraging feedback in AI models enables dynamic improvements in routing decisions and agent support, enhancing overall contact center performance.

E is incorrect: While managing agent workload is important, disabling feedback collection can lead to missed insights; better approaches include adjusting feedback timing or frequency.

F is incorrect: While anonymity can increase honesty, in Dynamics 365 Contact Center AI, linking feedback to specific agents is critical for targeted coaching and performance improvement.

A is correct: Assigning the agent to a routing skill profile with defined proficiency levels ensures that the work distribution system can match work items to agents based on skill requirements and agent capabilities.

B is incorrect: Setting 24/7 availability disregards the agent's actual working hours and can lead to inefficient routing and agent overload.

C is incorrect: Automatic escalation is a separate process and does not influence how work items are initially distributed to agents based on skills.

D is incorrect: While preferred communication channels help route work, without skill assignments, the system cannot prioritize work distribution based on agent proficiency.

A is correct: Enabling the pre-chat survey and mapping the fields ensures that customer responses are captured and stored in Dynamics 365 entities before the conversation is routed.

B is incorrect: Post-chat surveys collect feedback after the conversation, not before routing, so this does not configure the pre-chat survey.

C is incorrect: Pre-chat surveys must be completed before agent interaction; triggering after agent acceptance defeats the purpose.

D is incorrect: Disabling the pre-chat survey prevents capturing initial customer context and requirements, which is essential for efficient routing.

A is incorrect: Importing static Excel datasets does not support real-time updates as the data is fixed at the time of import and requires manual refresh.

B is correct: Direct live query connections enable KPIs to reflect real-time data changes without manual refresh, which is essential for dynamic contact center monitoring.

C is incorrect: Scheduled refreshes provide periodic updates but are not suitable for real-time KPI customization needed in contact center scenarios.

D is incorrect: Manual exports are inefficient and do not support dynamic or real-time KPI updates, reducing the responsiveness of supervisor analytics.

A is correct: Power BI Embedded allows embedding interactive reports within the workspace, providing real-time analytics without users leaving the environment.

B is correct: Properly assigning security roles is essential to control access, as reports often contain sensitive contact center data.

C is incorrect: Out-of-the-box reports should not have their dataset schema modified; customization is done via parameters or filters, not schema changes.

D is incorrect: Embedding reports requires authentication configuration to ensure secure access; direct iframe embedding without auth is not supported.

E is correct: The workspace settings allow adding report URLs and managing refresh schedules, ensuring data remains current and accessible.

F is incorrect: Disabling caching can degrade performance; best practice is to configure appropriate refresh intervals rather than disabling caching entirely.

A is incorrect: Round Robin assigns tasks in a fixed cyclic order without considering the current workload of agents.

B is correct: The Least Active method assigns tasks to the agent with the fewest active tasks, effectively balancing workload dynamically.

C is incorrect: Longest Idle assigns tasks to the agent who has been idle the longest, which may not reflect current workload balance.

D is incorrect: Random assignment distributes tasks arbitrarily and does not optimize for workload distribution.

A is incorrect: Relying on a single, unfiltered knowledge source can inundate the search with extraneous information, decreasing precision and negatively impacting the agent's experience.

B is correct: Role-based access controls ensure adherence to policies and relevance by limiting agents' visibility to only the knowledge content authorized and relevant to their specific roles.

C is correct: Activating semantic search employs AI to understand user intent and context, thereby significantly improving the relevance of results from external knowledge sources.

D is incorrect: Deactivating synonyms and stemming typically reduces search versatility and can adversely affect recall, leading to less effective search outcomes.

E is incorrect: Daily full reindexing can be resource-intensive and is often unnecessary; incremental updates are generally more efficient unless significant changes require a complete reindex.

F is correct: Incremental data refreshing minimizes latency and resource usage by updating only modified content, which guarantees efficient and current knowledge updates.