Fixing Acrolinx

Author: Randall-Knapp

agent-framework/user-guide/workflows/orchestrations/magentic

@@ -18,41 +18,41 @@ quiz:
     choices:
     - content: "Start connecting every line-of-business system to a new agent"
       isCorrect: false
-      explanation: "Incorrect. This step does not address governance and security, which are critical to avoiding agent sprawl and security drift."
+      explanation: "Incorrect. This step doesn't address governance and security, which are critical to avoiding agent sprawl and security drift."
     - content: "Define and enforce agent governance (roles, policies, development process) across teams"
       isCorrect: true
       explanation: "Correct. According to the module, CAF foundations should transition directly into governance and security for agents, with clear policies and process gates. This step ensures that agent development is controlled and aligned with organizational standards, reducing risks such as sprawl and security gaps."
-    - content: "Purchase additional GPU capacity"
+    - content: "Purchase more GPU capacity"
       isCorrect: false
-      explanation: "Incorrect. While GPU capacity may be important for agent performance, it does not address governance or security concerns."
+      explanation: "Incorrect. While GPU capacity may be important for agent performance, it doesn't address governance or security concerns."
     - content: "Skip to production and monitor later"
       isCorrect: false
       explanation: "Incorrect. Skipping governance and security steps increases the risk of agent sprawl and security drift."
   - content: "Which factor should a solution architect consider first when deciding whether to use a SaaS agent or build a custom agent?"
     choices:
     - content: "The availability of GPU clusters"
       isCorrect: false
-      explanation: "Incorrect. GPU availability is not the primary consideration when deciding between SaaS and custom agents."
+      explanation: "Incorrect. GPU availability isn't the primary consideration when deciding between SaaS and custom agents."
     - content: "Whether a SaaS agent meets functional requirements"
       isCorrect: true
       explanation: "Correct. The source material emphasizes a 'SaaS agent first' principle, where architects should begin by determining if a SaaS agent meets the functional requirements. If it does, it should be adopted to maximize value and minimize unnecessary customization efforts."
     - content: "The number of developers on the project"
       isCorrect: false
-      explanation: "Incorrect. The number of developers is not the primary factor in deciding between SaaS and custom agents."
+      explanation: "Incorrect. The number of developers isn't the primary factor in deciding between SaaS and custom agents."
     - content: "The preferred programming language"
       isCorrect: false
-      explanation: "Incorrect. Programming language preference is not the primary consideration when choosing between SaaS and custom agents."
+      explanation: "Incorrect. Programming language preference isn't the primary consideration when choosing between SaaS and custom agents."
   - content: "You're designing a solution that must handle confidential finance data and public product data, with different teams owning each and separate release cycles. Which architecture is most appropriate to start with?"
     choices:
     - content: "Single agent with broad permissions"
       isCorrect: false
-      explanation: "Incorrect. A single agent with broad permissions does not enforce separation of concerns or align with governance requirements."
+      explanation: "Incorrect. A single agent with broad permissions doesn't enforce separation of concerns or align with governance requirements."
     - content: "Multi-agent with isolated permissions and explicit interfaces"
       isCorrect: true
-      explanation: "Correct. The module advises that when solutions cross security or compliance boundaries and involve distinct team ownership, a modular multi-agent design with isolated permissions and explicit interfaces is preferred. This approach enforces separation of concerns and aligns with governance requirements."
+      explanation: "Correct. The module advises that when solutions cross security or compliance boundaries and involve distinct team ownership, a modular multi-agent design with isolated permissions (and explicit interfaces) is preferred. This approach enforces separation of concerns and aligns with governance requirements."
     - content: "Single agent using persona switching"
       isCorrect: false
-      explanation: "Incorrect. Persona switching does not provide the necessary separation of permissions and governance for handling confidential and public data."
+      explanation: "Incorrect. Persona switching doesn't provide the necessary separation of permissions and governance for handling confidential and public data."
     - content: "Single agent with larger context windows"
       isCorrect: false
-      explanation: "Incorrect. Larger context windows do not address the need for isolated permissions and explicit interfaces."
+      explanation: "Incorrect. Larger context windows don't address the need for isolated permissions and explicit interfaces."

learn-pr/wwl/design-overall-ai-strategy-business-solutions/18-knowledge-check.yml

@@ -16,7 +16,7 @@ By the end of this unit, learners will be able to:
 
 ## Understanding prompt engineering in business solutions
 
-Prompt engineering is the deliberate design of instructions that guide AI models to produce reliable outputs. Because AI systems do not understand intent, they rely entirely on the clarity, structure, and context provided in a prompt.
+Prompt engineering is the deliberate design of instructions that guide AI models to produce reliable outputs. Because AI systems don't understand intent, they rely entirely on the clarity, structure, and context provided in a prompt.
 
 In business environments, effective prompt engineering ensures:
 

learn-pr/wwl/design-overall-ai-strategy-business-solutions/includes/12-provide-prompt-engineering-guidelines-techniques.md

@@ -136,7 +136,7 @@ These personas collectively ensure that AI workloads are secure, high-performing
 
 ## Additional roles and considerations
 
-The above list of roles are not comprehensive and will vary greatly from organization to organization. The roles may be specialized and further broken out or consolidated depending on the size of an organization or project. It's highly recommended to conduct activities such as a role-mapping workshop, gap analysis exercise, and a RACI building activity to turn the list of roles into a practical and sustainable framework for the organizations. 
+The above list of roles aren't comprehensive and will vary greatly from organization to organization. The roles may be specialized and further broken out or consolidated depending on the size of an organization or project. It's highly recommended to conduct activities such as a role-mapping workshop, gap analysis exercise, and a RACI building activity to turn the list of roles into a practical and sustainable framework for the organizations. 
 
 Additional roles not listed may be discovered such as the below for example:
 

learn-pr/wwl/design-overall-ai-strategy-business-solutions/includes/13-identify-key-business-user-roles-ai-workloads.md

@@ -106,7 +106,7 @@ Your training program should teach users:
 
 ### How AI interprets prompts
 
-- AI does not infer intent
+- AI doesn't infer intent
 
 - Clarity and specificity matter
 

learn-pr/wwl/design-overall-ai-strategy-business-solutions/includes/17-design-user-prompt-training-ai-solution-adoption.md

@@ -27,35 +27,35 @@ By the end of this unit, learners will be able to:
 
 Design multi-agent systems by **assigning the right platform to each role**:
 
-- **Microsoft 365 Copilot (SaaS)** — Domain agents embedded in Microsoft 365 experiences (for example, summarization, drafting, scheduling) or . Use to **activate immediate value** where capabilities fit the task, accepting limited customization. 
+- **Microsoft 365 Copilot (SaaS)**—Domain agents embedded in Microsoft 365 experiences (for example, summarization, drafting, scheduling) or . Use to **activate immediate value** where capabilities fit the task, accepting limited customization. 
 
-- **Copilot Studio (low-code SaaS)** — Rapidly build **task and retrieval** agents with prebuilt connectors and guardrails; ideal for business-led processes, moderate customization, and quick iteration. 
+- **Copilot Studio (low-code SaaS)**—Rapidly build **task and retrieval** agents with prebuilt connectors and guardrails; ideal for business-led processes, moderate customization, and quick iteration. 
 
-- **Microsoft Foundry (pro-code)** — Build **connected agents** and sophisticated **multi-agent** workflows with deeper control over orchestration, tools, and runtime; best for strategic, high-integration scenarios
+- **Microsoft Foundry (pro-code)**—Build **connected agents** and sophisticated **multi-agent** workflows with deeper control over orchestration, tools, and runtime; best for strategic, high-integration scenarios
 
-- **Design guidance:** Start with SaaS agents where they meet functional requirements; introduce **Copilot Studio** for tailored workflows; escalate to **Foundry** for complex orchestration, custom tools, and code-first agents
+- **Design guidance:**—Start with SaaS agents where they meet functional requirements; introduce **Copilot Studio** for tailored workflows; escalate to **Foundry** for complex orchestration, custom tools, and code-first agents
 
 ## Orchestration patterns with the Microsoft Agent Framework
 
 When agents collaborate, adopt **explicit orchestration** rather than ad hoc chaining. The **Microsoft Agent Framework SDK** provides patterns you can mix and match: 
 
-**Sequential orchestration** — Deterministic pipeline for staged tasks (plan → enrich → verify → act). For more information, see [Sequential orchestration](/agent-framework/user-guide/workflows/orchestrations/sequential?pivots=programming-language-csharp).
+**Sequential orchestration**—Deterministic pipeline for staged tasks (plan → enrich → verify → act). For more information, see [Sequential orchestration](/agent-framework/user-guide/workflows/orchestrations/sequential?pivots=programming-language-csharp).
 
 :::image type="content" source="../media/input-flow-chart.png" alt-text="Diagram of sequential orchestration showing a deterministic pipeline where tasks flow from one agent to the next in order.":::
 
-**Concurrent orchestration** — Parallel agents tackle independent subtasks; aggregate and reconcile results. For more information, see [Concurrent orchestration](/agent-framework/user-guide/workflows/orchestrations/concurrent?pivots=programming-language-csharp).
+**Concurrent orchestration**—Parallel agents tackle independent subtasks; aggregate and reconcile results. For more information, see [Concurrent orchestration](/agent-framework/user-guide/workflows/orchestrations/concurrent?pivots=programming-language-csharp).
 
 :::image type="content" source="../media/concurrent-orchestration.png" alt-text="Diagram of concurrent orchestration showing parallel agents handling independent subtasks simultaneously before aggregating results.":::
 
-**Group chat orchestration** — A mediated conversation where agents contribute proposals and a moderator agent arbitrates. For more information, see [Group chat orchestration](/agent-framework/user-guide/workflows/orchestrations/group-chat?pivots=programming-language-csharp).
+**Group chat orchestration**—A mediated conversation where agents contribute proposals and a moderator agent arbitrates. For more information, see [Group chat orchestration](/agent-framework/user-guide/workflows/orchestrations/group-chat?pivots=programming-language-csharp).
 
 :::image type="content" source="../media/group-chat-orchestration.png" alt-text="Diagram of group chat orchestration showing multiple agents contributing proposals in a mediated conversation with a moderator agent.":::
 
-**Handoff orchestration** — Transfer context and control to a specialist agent (or a human) when a threshold or rule triggers escalation. For more information, see [Handoff orchestration](/agent-framework/user-guide/workflows/orchestrations/handoff?pivots=programming-language-csharp).
+**Handoff orchestration**—Transfer context and control to a specialist agent (or a human) when a threshold or rule triggers escalation. For more information, see [Handoff orchestration](/agent-framework/user-guide/workflows/orchestrations/handoff?pivots=programming-language-csharp).
 
 :::image type="content" source="../media/handoff-orchestration.png" alt-text="Diagram of handoff orchestration showing context and control transferring from one agent to a specialist agent or human when an escalation threshold is reached.":::
 
-**Magentic orchestration** — Pattern for dynamic specialization where a "magnet" pulls in the right expert agents at runtime. For more information, see [Magentic orchestration](/agent-framework/user-guide/workflows/orchestrations/magentic?pivots=programming-language-csharp).
+**Magentic orchestration**—Pattern for dynamic specialization where a "magnet" pulls in the right expert agents at runtime. For more information, see [Magentic orchestration](/agent-framework/user-guide/workflows/orchestrations/magentic?pivots=programming-language-csharp).
 
 :::image type="content" source="../media/magnetic-orchestration.png" alt-text="Diagram of Magentic orchestration showing a central orchestrator dynamically pulling in specialized expert agents at runtime based on task requirements.":::
 
@@ -99,7 +99,7 @@ A robust multi-agent solution in **Foundry** follows these steps:
 |---|---|---|---|
 | **Domain assistant (productivity)** | Microsoft 365 Copilot | Immediate value inflow of work | Handoff / group chat |
 | **Business workflow agent** | Copilot Studio | Rapid iteration, connectors, guardrails | Sequential / handoff |
-| **Integration/orchestration agent** | Foundry | Pro-code tools, complex flows, custom evals | Concurrent / sequential / Magentic |
+| **Integration/orchestration agent** | Foundry | Pro-code tools, complex flows, custom evaluations | Concurrent / sequential / Magentic |
 
 ## Security, governance, and operability considerations
 
@@ -115,7 +115,7 @@ A robust multi-agent solution in **Foundry** follows these steps:
 
 - Instrument **handoff latency, tool failure rate, and decision quality**; log messages and artifacts for audits.
 
-### Rollback & human-in-the-loop
+### Roll back & human-in-the-loop
 
 - Gate high-risk actions with approvals; design "break-glass" paths.
 

learn-pr/wwl/design-overall-ai-strategy-business-solutions/includes/4-design-multi-agent-solution.md

@@ -1,6 +1,6 @@
 ## Unit overview
 
-This unit equips solution architects with expert-level skills to analyze, map, and design high-value business use cases for **prebuilt Microsoft 365 Copilot agents**. Prebuilt agents accelerate productivity by streamlining routine tasks, enabling fast information retrieval, and providing structured guidance across business processes. Prebuilt agents deliver value quickly because they do not require custom development, yet remain customizable through organizational knowledge and configured behaviors.
+This unit equips solution architects with expert-level skills to analyze, map, and design high-value business use cases for **prebuilt Microsoft 365 Copilot agents**. Prebuilt agents accelerate productivity by streamlining routine tasks, enabling fast information retrieval, and providing structured guidance across business processes. Prebuilt agents deliver value quickly because they don't require custom development, yet remain customizable through organizational knowledge and configured behaviors.
 
 This unit synthesizes foundational principles of AI agents, core use case patterns, and applied scenario mapping techniques used by architects to translate business needs into effective agent-powered solutions.
 

learn-pr/wwl/design-overall-ai-strategy-business-solutions/includes/5-develop-use-cases-prebuilt-agents-solution.md

@@ -36,7 +36,7 @@ Extensions typically include:
 
 Build custom agents when:
 
-- You need specialized workflows that Copilot cannot handle
+- You need specialized workflows that Copilot can't handle
 
 - Your scenario requires custom reasoning patterns, multi-step logic, or orchestration
 

learn-pr/wwl/design-overall-ai-strategy-business-solutions/includes/8-determine-build-custom-agents-extend-copilot.md

@@ -2,19 +2,19 @@
 
 This unit equips solution architects with the expert-level skills needed to determine when organizations should invest in building **custom AI models**, instead of relying on prebuilt or catalog models available through platforms such as **Microsoft 365 Copilot**, **Microsoft Foundry**, or **Azure OpenAI model catalog**.
 
-You'll evaluate business, technical, operational, and cost-efficiency drivers required to justify custom model development.
+You evaluate business, technical, operational, and cost-efficiency drivers required to justify custom model development.
 
 ## Understanding the decision landscape
 
 Choosing whether to build a custom model is a strategic decision with major implications for **cost, time-to-market, maintainability, security, and talent requirements**.
 
 In general, organizations should create custom AI models only when:
 
-- The business problem cannot be solved accurately with existing pre-trained or fine-tuned model
+- The business problem can't be solved accurately with existing pretrained or fine-tuned model
 
 - **Domain specificity**, **sensitive workflows**, or **high-impact decisioning** demands deeper customization
 
-- **Model behavior must be highly predictable or governed**, and prebuilt models cannot meet compliance thresholds
+- **Model behavior must be highly predictable or governed**, and prebuilt models can't meet compliance thresholds
 
 - **ROI modeling** confirms long-term efficiency gains that outweigh higher upfront engineering cost
 
@@ -34,11 +34,11 @@ Prebuilt or catalog models fit well when:
 
 - The agent interacts primarily with **enterprise knowledge sources**, not highly specialized reasoning tasks
 
-- The domain data is **not complex** or does not require deep contextual understanding
+- The domain data is **not complex** or doesn't require deep contextual understanding
 
 - **Time-to-value** is a priority
 
-- Teams want **low-cost**, low-risk deployment
+- A team wants **low-cost**, low-risk deployment
 
 ### Examples
 
@@ -60,7 +60,7 @@ A custom model is recommended in the following situations.
 
 ### Domain-specific intelligence is required
 
-If the system must understand organization-specific language, terminology, processes, or industry technical vocabulary, a pre-trained model may lack the accuracy needed.
+If the system must understand organization-specific language, terminology, processes, or industry technical vocabulary, a pretrained model may lack the accuracy needed.
 
 #### Indicators
 
@@ -134,7 +134,7 @@ Custom models are cost-effective when:
 
 - Performance gains will directly translate to revenue lift or efficiency gain
 
-This aligns with Azure training on **cost-efficiency decision-making**.
+This data aligns with Azure training on **cost-efficiency decision-making**.
 
 ### Multi-agent systems requiring custom reasoning
 
@@ -168,7 +168,7 @@ Custom model development requires:
 
 - Access to skilled data scientists and MLOps engineers
 
-If these requirements are not yet met, extending Microsoft 365 Copilot is often the better starting point.
+If these requirements aren't yet met, extending Microsoft 365 Copilot is often the better starting point.
 
 ## Architecture decision framework (text-based visual)