AI security

Nudge Security now discovers AI agents your team builds on Datadog and HubSpot, right from the browser.

​

Every agent flows into your AI agents inventory under Identities, where you'll see its creator, components, risk insights, and more.

​

This is available as part of the AI agent discovery research preview. Ensure you have browser-based agent discovery turned on in your browser extension settings.

​

Not yet in the research preview? Request access from the AI Agents tab under Identities, or start a 14-day free trial if you're new to Nudge Security.

You can now catch sensitive data before it reaches an AI tool.

​

When the browser extension detects sensitive data like PII, secrets, or financial info in an AI conversation, an in-browser nudge shows the employee exactly what was found and lets them redact it in one click, or justify sending it. Optionally turn on enforcement and redaction becomes the only way to send, so unredacted sensitive data never leaves the browser.

​

Head to Settings > In-browser nudges to turn on sensitive data nudges and choose whether to enforce redaction.

​

New to Nudge Security? Start a 14-day free trial.

AI conversation monitoring now detects sensitive data shared in Amazon Quick Suite, Amazon's agentic AI platform. When someone enters secrets, PII, financial data, or health data into an Amazon Quick chat, Nudge Security flags it.

​

Configure detection in your Browser Extension Settings. To learn more, see AI conversation monitoring.

​

New to Nudge Security? Start a 14-day free trial to detect and respond to AI data risks.

Nudge Security now discovers AI agents your team builds on Amazon Quick, AWS's agentic AI platform, right from the browser.

​

Every agent flows into your AI agents inventory under Identities, where you'll see its creator, components, risk insights, and more.

​

This is available as part of the AI agent discovery research preview. Ensure you have browser-based agent discovery turned on in your browser extension settings.

​

Not yet in the research preview? Request access from the AI Agents tab under Identities, or start a 14-day free trial if you're new to Nudge Security.

Nudge Security calculates inherent and residual risk scores for every SaaS and AI vendor in your environment, mapped to a Low, Medium, High, or Critical level and updated continuously as risk factors change. These new scores complement business criticality tiers, which signal potential impact rather than current risk.

‍

Inherent risk reflects your risk before accounting for any controls you've put into place, whereas residual risk includes those controls. Nudge Security provides a transparent breakdown of the factors driving each score, including risk factors associated with each vendor's own security posture, internal risk factors based on your organization’s deployment and usage, and compensating controls you have in place to mitigate risk. With these scores, which are now in open beta, you can:

• Get risk context for the long tail of apps other tools miss. Every SaaS and AI app gets a score on Day One, no vendor cooperation required.
• Account for internal risk factors. Capture easily-overlooked risk factors like critical downstream connections and AI agents that can act through the app.
• Continuously monitor risk. Instead of a one-time snapshot, scores recalculate automatically as risk factors change, from new MCP server connections to vendor breach disclosures.
• Know what to fix next. See recommended actions ranked by projected residual risk reduction.

View and filter risk scores within your app inventory, or head to any app overview and click the Risk tab to see risk scores and recommended actions.

‍

New to Nudge Security? Start a 14-day free trial to see your real vendor risk exposure.

‍

Now, Nudge Security helps you triage and prioritize apps at the scale of modern SaaS and AI adoption by automatically assigning each app a business criticality tier (Low, Medium, High, or Critical) based on the data it typically handles.

‍

An app’s data access represents what’s at stake if the app is compromised. Nudge Security uses a proprietary model to infer the data types each AI or SaaS vendor typically handles. Next, each data type is classified by sensitivity, and business criticality is set based on the highest data sensitivity tier. With this update, which is now in open beta, you can:

• Automatically triage every app in your environment based on its potential impact to your business.
• Filter your inventory by 27 distinct data types, with no manual research required.
• Tailor data sensitivity tiers to match your organization's data governance model (Settings > Risk). Business criticality recalculates automatically, except where you've set a tier manually.
• Surface certain security posture findings only for Critical and High tier apps.

View business criticality tiers and data types within your app inventory, or within any app overview. Note: Previously tagged data types aren't lost. They now live under Data Types (Legacy), separate from the new, automatically populated Data Types field.

‍

New to Nudge Security? Start a 14-day free trial to see what data your apps can access.

Nudge Security now discovers AI agents employees build across platforms that don't expose a public API. The Nudge Security browser extension inventories agents as employees create them across these platforms:

• Cursor automations
• Retool
• Zapier Agents
• ChatGPT workspace agents
• OpenAI Agents Workflows
• Atlassian Rovo (Jira, Bitbucket)
• Zoom AI Workflows
• HyperAgent

Every agent flows into your AI agents inventory under Identities, where you'll see its creator, components, risk insights, and more.

​

We're releasing this as part of the existing AI agent discovery research preview. If you're already participating in the research preview and have the Nudge Security browser extension deployed, head to your browser extension settings and toggle on browser-based agent discovery.

​​

Not yet in the research preview? You can request access from the AI Agents tab under Identities in the product navigation or start a 14-day free trial if you're new to Nudge Security.

AI agent discovery now supports Claude managed agents and LangSmith Fleet. Connect either platform to see who built each agent, what it's connected to, and what it's authorized to do.

​

Head to Settings > Connected Apps to add Claude managed agents or LangSmith Fleet.

​

To learn more about AI agent discovery, read our blog.

​

New to Nudge Security? Start a 14-day free trial to discover and inventory AI agents across your org.

Nudge Security now discovers and assesses the risk of AI agents across connected apps, including ChatGPT, Microsoft Copilot, Gemini, Salesforce Agentforce, ServiceNow, n8n, Tines, Abacus, and Workato.

‍

AI agent discovery is built on Nudge Security's connected apps infrastructure. Request access to get started.

‍

With it you can:

‍

• See agents across connected apps, including who built it, what it's connected to, and what it's authorized to do.
• Hold someone accountable for every agent. Assign a technical contact so each agent has someone responsible for it, and you always know who to go to when something needs to be addressed.
• Surface agentic risks, including publicly accessible agents, hardcoded credentials, unauthenticated MCP connections, agents connected to high-risk apps, and more.
• Engage creators directly, not just flag risks. When an agent needs attention, nudge the responsible contact, asking them to confirm its purpose, justify use, and remediate any risks.

‍

To discover agents across ChatGPT, Microsoft Copilot, Gemini, and the other supported connected apps, connect them from your connected apps settings. Read the blog to learn more.

‍

New to Nudge Security? Start a 14-day free trial to request access to AI agent discovery.

‍

‍

You can now map file uploads and copy/paste actions from your SaaS applications directly to AI conversation tools with a new card in the AI Usage Dashboard.

‍

This interactive chart allows you to:

• Spot high-volume data movements and potential risks at a glance.
• Rank flows by data volume, file count, or user count.
• Identify which AI tools are consuming the most corporate data.

Head to the AI Usage Dashboard to pinpoint where your corporate data is flowing and understand the blast radius of data shared with AI conversation tools.‍

‍

Note: The Nudge Security browser extension is required to populate this chart.

‍

We’ve expanded our OAuth risk insights to cover remote Model Context Protocol (MCP) connections.

‍

Discover remote MCP connections powering AI agents and tools that access your data, visualize their scopes and permissions, and identify sensitive "data highways" to assess risk across your AI workforce.

‍

You can now create notifications to alert your team when sensitive data, like PII, secrets, or financial info, is shared in AI conversations. Configure alerts based on the type of data detected or at the app level, and route them to Slack, Microsoft Teams, email, or a webhook so you can investigate risks faster.

Our AI conversation monitoring now detects when sensitive data is shared in conversations with AI chatbot tools. Configure monitoring for popular tools like ChatGPT, Claude, Gemini, Groq, Meta AI, Copilot, and more to identify when employees share secrets and credentials, PII, financial data, and healthcare data.

‍

This feature helps organizations protect their data and understand how employees engage with AI tools. View full conversation details at the user, account, and app level, with insights into:

• Sensitive data detections broken down by category
• Copy/paste activity with source detection
• File upload metadata (name, source, size, and type)
• Conversation metrics and timestamps

‍

Configure detection settings and data storage preferences in your Browser Extension Settings.

‍

You can now easily deploy the browser extension on AI browsers like ChatGPT Atlas, Dia, and Perplexity's Comet, with dedicated deployment support. These additions allow organizations to secure employees everywhere work is happening, including in AI browsers.

‍

Head to your Nudge Security settings to deploy across your organization today.

‍

New to Nudge Security? Start a 14-day free trial to expand your SaaS and AI security coverage.

You can now set up a nudge to deliver your AI Acceptable Use Policy (AUP) directly in the browser when employees sign up for or log in to AI tools. The browser nudge will trigger for apps with an approved, acceptable, or no-approval status. This makes it easy to educate users in the moment, reinforce policy awareness, and help teams use AI safely across your organization.

‍

Browser nudging is only available when you deploy our browser extension. Get deployed now for the most comprehensive coverage.

Body: Our new card in the AI Usage dashboard now shows daily active users across (DAUs) across AI chatbot tools. You can also see this at the app, account, and individual user level on their detail pages, giving you visibility into how employees are using AI. Track adoption trends, overall engagement, and identify opportunities to optimize AI tool usage across your organization.

‍

Available for organizations using the Nudge Security browser extension. Deploy now.

We’ve refreshed our AI Acceptable Use Policy (AUP) playbook to make governance easier and more actionable. Existing customers using the previous playbook don’t need to take any action unless they want to update their policy.

‍

With this updated playbook, IT admins can:

‍

• Create, update, and manage AI policies directly in the product.
• Deliver policies over Slack or Microsoft Teams to streamline enforcement efforts and reduce manual work.
• Set up nudges to automatically send policies to employees the first time they sign up for an AI tool.
• Track AUP acceptance across the workforce in the AI usage dashboard, Users table, and user details page.

‍

This update helps organizations strengthen trust in AI, automate AI governance at scale, and give leaders clear visibility into policy adoption.

We’ve added a new card to the security tab of the app details page. This new card summarizes each app’s AI data training policy, including whether your data is used for training, available opt-out options, retention periods, and other relevant information. This makes it easier for teams to evaluate SaaS and AI tools by showing how each app handles data without requiring a review of lengthy documentation.

Our new AI governance playbook guides you through evaluating and categorizing AI tools discovered in your estate. The playbook helps you configure rules and policies that align with your developing governance framework. During this workflow you can:

• Review unapproved AI applications
• Remove not permitted AI applications
• Revoke unnecessary access permissions
• Establish guardrails for managing new AI applications as they’re introduced

Nudge Security now shows you which AI tools have access to sensitive data like email, files, and source code in the AI usage dashboard. You can easily see this information by department and change app status directly in the dashboard, helping you reduce the risk of sensitive data exposure to AI tools.