Key Takeaways
-
Work has moved to the interaction layer, creating massive in-session vulnerabilities and fundamentally breaking traditional security models.
-
In this video, you’ll see how Akamai enables real-time visibility and adaptive enforcement at the interaction layer without changing browsers, workflows, or network architecture.
Video transcript
Video transcript
Hey, let's take a quick look at how LayerX helps organizations stay secure in the age of AI without getting in the way of productivity. LayerX is a security platform that gives security teams full visibility and control over AI interactions across browsers, desktop apps, IDEs, on-device agents, and browser extensions, all without touching your network architecture or slowing anyone down.
There's a problem that's happening everywhere these days: users are pasting data into AI tools like ChatGPT, Copilot, or Gemini, often without realizing it includes sensitive information like PII, customer records, or financial data. Most organizations have absolutely no visibility into that, let alone any control over it.
As an example, we set a policy in LayerX that blocks Social Security numbers from being pasted into GenAI tools. When pasting fake customer data with a Social Security number into ChatGPT, the sensitive part gets automatically redacted. We get an alert, and nothing risky gets through. We can still use the tool; it just works in a way that's secure.
LayerX keeps sensitive data from leaking out while keeping people productive. The same idea applies to file uploads. In Claude, trying to upload a file with source code is identified and blocked based on the set policy. You can apply this control to any site, any action—uploads, downloads, copy, paste—through the browser or desktop apps.
LayerX can also block access to AI tools not on your approved list. For example, attempting to access DeepSeek results in an automatic redirect to an approved tool like ChatGPT. No shadow AI, no workarounds, just employees working within vetted tools.
Who the user is matters a lot in AI security. Identity determines if they are on a free or business account, data retention periods, whether data is used for model training, and if corporate security policies apply. LayerX enforces controls based on identity, corporate versus personal accounts, user groups, roles, and login context.
For example, pasting content into Gemini using a personal Gmail account is blocked. Switching to a corporate account allows the same action. LayerX also controls which identity is logged into any AI tool and can enforce adaptive policies. Guest mode or personal email logins can be blocked to prevent logins from non-corporate identities or those without SSO.
LayerX also secures browser and IDE extensions. It can automatically block risky extensions, like a DeepSeek AI extension, at the moment of installation. Users cannot enable blocked extensions even in the browser's management page. LayerX analyzes, categorizes, and assesses the risk of every extension based on risk level and identity. Extensions can also be restricted to specific sites; for instance, the Claude Chrome extension can work fine generally but be blocked on Salesforce to protect sensitive CRM data.
As AI agents become more autonomous, LayerX identifies agentic activity and applies guardrails. Using an AI browser like Perplexity Comet to copy text into ChatGPT via agentic mode is blocked because LayerX identifies the action was initiated by an AI agent, not a human.
LayerX protects users from risky websites and phishing attacks using a built-in AI-based risk engine. It can automatically scan and block zero-day phishing sites from sources like PhishTank without requiring user action.
Everything is configurable through the policy engine. Policies can be built based on websites, applications, and user activities like copy-paste, uploads, downloads, text input, login events, and AI prompts. Enforcement options include monitoring, warning users, restricting with or without bypass, watermarking, redirecting traffic, redacting information, or taking screenshots.
Triggered policies are logged in the alert section, which can be viewed in the console or sent to a SIEM for analysis. LayerX provides full visibility into all AI and web interactions, including accounts used, full conversations, shared files, and extension events. It is a horizontal security platform for AI, web, and SaaS activity.
LayerX is deployed via a browser extension (which works in incognito and AI browsers) or a desktop agent (covering local AI apps and IDEs). It is lightweight with no performance hits.
LayerX covers AI usage control, SaaS security, DLP, browser security, BYOD, and more. Visit layerxsecurity.com to schedule a demo. Thank you for watching and let's make security work for your business, not against it.