Small fixes (contractions, formatting)

Author: pedrosousa

src/content/docs/learning-paths/holistic-ai-security/concepts/shadow-ai.mdx

@@ -9,7 +9,7 @@ Shadow AI is the use of any AI tool or application by employees without the form
 
 ## Manage the risks of shadow AI
 
-Since it may not be feasible to eliminate all instances of shadow AI, organizations can manage the risks by taking a balanced approach. First, it's essential to foster collaboration and open communication between IT and other departments to understand why employees are using these tools. Next, implement guardrails by developing a flexible governance framework with clear guidelines on which types of AI systems can be used and how sensitive information should be handled. Finally, monitor AI usage with network tools to track what's being used and educate employees on the risks of unauthorized AI. 
+Since it may not be feasible to eliminate all instances of shadow AI, organizations can manage the risks by taking a balanced approach. First, it is essential to foster collaboration and open communication between IT and other departments to understand why employees are using these tools. Next, implement guardrails by developing a flexible governance framework with clear guidelines on which types of AI systems can be used and how sensitive information should be handled. Finally, monitor AI usage with network tools to track what is being used and educate employees on the risks of unauthorized AI.
 
 To combat the risks associated with shadow AI, Cloudflare's API-driven CASB integrates with SaaS applications and cloud environments to scan for misconfigurations, unauthorized user activity, shadow IT, and other data security issues that can occur after a user has successfully logged in.
 

src/content/docs/learning-paths/holistic-ai-security/get-started/additional-setup.mdx

@@ -6,7 +6,7 @@ sidebar:
   order: 1
 ---
 
-To use the recommended AI security solutions in this learning path, you'll need to complete some additional setup for your Zero Trust deployment, such as turning on [Shadow IT Discovery](/cloudflare-one/insights/analytics/shadow-it-discovery/#turn-on-shadow-it-discovery).
+To use the recommended AI security solutions in this learning path, you will need to complete some additional setup for your Zero Trust deployment, such as turning on [Shadow IT Discovery](/cloudflare-one/insights/analytics/shadow-it-discovery/#turn-on-shadow-it-discovery).
 
 :::note
 If you are new to securing your Internet traffic, refer to [Secure your internet traffic and SaaS apps](/learning-paths/secure-internet-traffic/concepts/) to get started and then return to this learning path to continue securing your organizations.

src/content/docs/learning-paths/holistic-ai-security/get-started/define-ai-risk-tolerance.mdx

@@ -6,12 +6,11 @@ sidebar:
   order: 3
 ---
 
-Before you can secure AI, you need to understand your organization's goals and concerns for your users and data. This first step is crucial for anyone new to AI security because it helps you align on your overall risk tolerance for generative AI. 
+Before you can secure AI, you need to understand your organization's goals and concerns for your users and data. This first step is crucial for anyone new to AI security because it helps you align on your overall risk tolerance for generative AI.
 
-The level of risk you're willing to accept will determine how quickly you need to act and how deeply you need to implement security measures. Your specific concerns will then guide the exact security approaches you take to monitor and control AI usage.
+The level of risk you are willing to accept will determine how quickly you need to act and how deeply you need to implement security measures. Your specific concerns will then guide the exact security approaches you take to monitor and control AI usage.
 
 To begin, consider these questions to help shape your strategy:
-- Sanctioned vs. shadow AI: Are your employees already using approved AI services, or are you primarily concerned with shadow AI — unapproved applications that may pose a risk? Note: Some approved vendors may offer embedded AI functionality. Consider whether those AI services are a risk, too.
+- Sanctioned versus shadow AI: Are your employees already using approved AI services, or are you primarily concerned with shadow AI — unapproved applications that may pose a risk? Note: Some approved vendors may offer embedded AI functionality. Consider whether those AI services are a risk, too.
 - Data sensitivity: What types of sensitive data do you currently monitor, classify, and protect? This will help you identify what you need to safeguard when employees interact with AI. If you have not determined this yet, refer to [Define your sensitive data](/learning-paths/secure-internet-traffic/build-http-policies/data-loss-prevention/#define-your-sensitive-data) to learn more.
 - Encouraging user behavior: Do you want to encourage or limit user engagement with AI? How much do you expect your employees to use AI in a controlled, sanctioned way?
----

src/content/docs/learning-paths/holistic-ai-security/monitor-ai-use/index.mdx

@@ -6,9 +6,9 @@ sidebar:
   order: 3
 ---
 
-Monitoring AI usage is one of the first steps in assessing holistic risk. Cloudflare provides a number of ways to understand both sanctioned and unsanctioned use of AI and chat applications within your Zero Trust deployment. Reviewing your detected usage should give you a sense of which AI apps are used in your organization as well as what type of policies you need to create based on what should be supported vs. restricted.
+Monitoring AI usage is one of the first steps in assessing holistic risk. Cloudflare provides a number of ways to understand both sanctioned and unsanctioned use of AI and chat applications within your Zero Trust deployment. Reviewing your detected usage should give you a sense of which AI apps are used in your organization as well as what type of policies you need to create based on what should be supported versus restricted.
 
 After you begin to understand your organization's AI usage, you should develop a balanced perspective on what constitutes sanctioned use of a tool. A good practice is to review applications in phases and then apply security policies based on those phases. For example, you can apply different security policies to applications that are considered approved, unapproved, unreviewed, or in review until they pass your corporate security standards and data privacy reviews.
 
-This section will detail options for assessing and understanding that risk by using inline (client) and out-of-band (API CASB) based methods. 
+This section will detail options for assessing and understanding that risk by using inline (client) and out-of-band (API CASB) based methods.
 

src/content/docs/learning-paths/holistic-ai-security/monitor-ai-use/monitor-prompts-responses.mdx

@@ -6,11 +6,11 @@ sidebar:
   order: 4
 ---
 
-When you enable [TLS decryption](/cloudflare-one/policies/gateway/http-policies/tls-decryption/#turn-on-tls-decryption) , you can review the prompts and responses for supported AI applications. This allows you to understand three key things about AI application usage:
-- The sanctioned and unsanctioned AI tools your users are engaging with .
+When you enable [TLS decryption](/cloudflare-one/policies/gateway/http-policies/tls-decryption/#turn-on-tls-decryption), you can review the prompts and responses for supported AI applications. This allows you to understand three key things about AI application usage:
+- The sanctioned and unsanctioned AI tools your users are engaging with.
 - How they are interacting with them.
 - What information they are sharing.
 
 ![Log entry for a prompt detected using AI prompt protection.](~/assets/images/learning-paths/holistic-ai-security/gateway-prompt-log.png)
 
-You can use this in conjunction with [DLP profiles](/cloudflare-one/policies/data-loss-prevention/dlp-profiles/) to detect sensitive data potentially being used in prompts, with or without explicitly blocking the action. You can use DLP to log AI prompt topics by turning on [Capture generative AI prompt content in logs](/cloudflare-one/policies/data-loss-prevention/dlp-policies/logging-options/#turn-on-ai-prompt-content-logging-for-a-dlp-policy) for the policy. 
+You can use this in conjunction with [DLP profiles](/cloudflare-one/policies/data-loss-prevention/dlp-profiles/) to detect sensitive data potentially being used in prompts, with or without explicitly blocking the action. You can use DLP to log AI prompt topics by turning on [Capture generative AI prompt content in logs](/cloudflare-one/policies/data-loss-prevention/dlp-policies/logging-options/#turn-on-ai-prompt-content-logging-for-a-dlp-policy) for the policy.

src/content/docs/learning-paths/holistic-ai-security/monitor-ai-use/review-inline-ai-use.mdx

@@ -8,13 +8,13 @@ sidebar:
 
 To review how your organization uses AI apps, start with the Cloudflare [Shadow IT dashboard](/cloudflare-one/insights/analytics/shadow-it-discovery/). You can filter the dashboard by application type and sort by Artificial Intelligence, which allows you to take several valuable actions, such as:
 
-Reviewing application usage: See which applications your users are engaging with and how often, based on traffic from connected devices or sites.
-Accessing application details: Learn more about the applications in use by reviewing their entries in the App Library.
-Designating application status: Use a [status marker](/cloudflare-one/insights/analytics/shadow-it-discovery/#approval-status) to categorize applications, so you can apply specific security policies to them.
+- Reviewing application usage: See which applications your users are engaging with and how often, based on traffic from connected devices or sites.
+- Accessing application details: Learn more about the applications in use by reviewing their entries in the App Library.
+- Designating application status: Use a [status marker](/cloudflare-one/insights/analytics/shadow-it-discovery/#approval-status) to categorize applications, so you can apply specific security policies to them.
 
 ![Shadow IT dashboard showing utilization of application statuses (Approved, Unapproved, In Review, Unreviewed).](~/assets/images/learning-paths/holistic-ai-security/shadowit-dashboard-ai-apps.png)
 
 This approach is most relevant for applications that your organization has not yet sanctioned. For example, if your business uses Google Workspace and its primary AI tool is Gemini, you would likely mark Gemini as *Approved*. While you can configure more specific data loss policies later, you can also target approved applications with greater specificity to detect risks unique to your organization.
 
-![The App Library filtered to show some of the supported Artificial Intelligence applications..](~/assets/images/learning-paths/holistic-ai-security/ai-app-type-app-library.png)
+![The App Library filtered to show some of the supported Artificial Intelligence applications.](~/assets/images/learning-paths/holistic-ai-security/ai-app-type-app-library.png)
 

src/content/docs/learning-paths/holistic-ai-security/monitor-ai-use/review-out-of-band-ai.mdx

@@ -6,6 +6,6 @@ sidebar:
   order: 3
 ---
 
-If your organization does not use the Cloudflare device client, or does not proxy HTTP traffic, you can still get valuable data about shadow AI usage if you use the Google Workspace, Microsoft 365, or Github integrations for the Cloudflare Cloud Access Security Broker (CASB).
+If your organization does not use the Cloudflare device client, or does not proxy HTTP traffic, you can still get valuable data about shadow AI usage if you use the Google Workspace, Microsoft 365, or GitHub integrations for the Cloudflare Cloud Access Security Broker (CASB).
 
-The CASB provides detailed information about your SaaS environment, including changes to sensitive data, content, and application settings. It works even if your users don't have the Cloudflare device client installed. By using CASB integrations with your core Single Sign-On (SSO) provider, you can see if users have authenticated to any third-party applications. This offers a clear, non-invasive way to understand tool usage across your organization without needing to deploy a client.
+The CASB provides detailed information about your SaaS environment, including changes to sensitive data, content, and application settings. It works even if your users do not have the Cloudflare device client installed. By using CASB integrations with your core Single Sign-On (SSO) provider, you can see if users have authenticated to any third-party applications. This offers a clear, non-invasive way to understand tool usage across your organization without needing to deploy a client.

src/content/docs/learning-paths/holistic-ai-security/secure-approved-ai-models-tools/index.mdx

@@ -12,7 +12,7 @@ As you decide which AI tools to sanction within your organization, you can devel
 
 Cloudflare supports API-based cloud access security broker (CASB) integrations with popular AI services like OpenAI (ChatGPT), Anthropic (Claude), and Google Gemini. These integrations allow you to report on AI tool usage and flag sensitive data with Data Loss Prevention (DLP) Profiles.
 
-Since these integrations are out-of-bound connections to SaaS applications, they don't require inline user traffic to create detections. This means you can immediately gain visibility into how your employees are using sanctioned AI tools without having to install the Cloudflare device client on every user's machine.
+Since these integrations are out-of-bound connections to SaaS applications, they do not require inline user traffic to create detections. This means you can immediately gain visibility into how your employees are using sanctioned AI tools without having to install the Cloudflare device client on every user's machine.
 
 ![OpenAI API CASB Integration showing riskier features that are toggled on, security posture risks like unused admin credentials, and an uploaded attachment with a DLP profile match.](~/assets/images/learning-paths/holistic-ai-security/casb-posture-findings-openai.png)
 
@@ -32,7 +32,7 @@ The Model Context Protocol (MCP) is an emerging standard that allows AI agents t
 
 Cloudflare has been an early supporter of the MCP standard. Many of our customers are already building custom MCP servers and use cases, and our engineering teams have worked to deliver MCP functionality for our public API. You can review our [changelog](/changelog/?product=ai) to see some of the MCP servers we have already released.
 
-Just like an API, an MCP server is a primary entry point for AI agents to interact with and manipulate your structured data. Since anyone can build and host an MCP server, it's crucial to have a comprehensive secure access strategy as your business starts to adopt these new agentic workflows.
+Just like an API, an MCP server is a primary entry point for AI agents to interact with and manipulate your structured data. Since anyone can build and host an MCP server, it is crucial to have a comprehensive secure access strategy as your business starts to adopt these new agentic workflows.
 
 As with any new technology, it can be difficult to manage the growing use of MCP servers among your employees. This space will continue to expand, but your first step should always be to gain visibility and understand usage before developing a security strategy to control inputs and outputs.