Aikido Noon Explained: AI Governance, Pricing and Early Access

Aikido announced Noon on October 20, a workforce security platform combining AI governance, email and device protection, and attack simulations. The useful change for AI teams is visibility beyond their own applications: Noon is designed to govern the tools, models and MCP connections employees use at work.

The public entry point is still early access, not an unrestricted general release. Its published plans include a free tier for up to ten users and a Pro tier listed at $7 per user per month.

What Aikido Noon adds

Aikido's existing focus is software and developer security. Its Noon announcement extends that scope to the wider workforce, bringing together email filtering, endpoint controls, AI-use policies and employee training.

The company says Noon can discover models, AI tools and MCP connections, enforce team-specific policies and identify sensitive information entering prompts. Training includes simulated email, voice, SMS and deepfake attacks. Those are product claims, not independent evidence of detection accuracy.

Noon also incorporates the teams behind Cyberly and Outkept, which Aikido acquired in September. The new announcement is the combined workforce product, rather than a newly released foundation model or another coding assistant.

How Noon AI governance works

The AI governance product page describes an extension deployed manually or through mobile device management. It lists macOS, Windows and Linux, with policies applied at company, team or individual level.

Administrators can categorize tools as allowed, disallowed or requiring approval, and set policies for models and MCP servers. The dashboard is intended to show which tools employees use and what those tools connect to.

MCP, or Model Context Protocol, gives AI applications a standardized way to connect to tools and data. In practical terms, approving a chatbot and approving every connection it can make are different decisions. A model used only to rewrite public copy has a different exposure from the same model connected to internal files.

Consider a small engineering team that already approves one coding assistant. An employee adds a remote MCP server to make that assistant more useful. The important question becomes what the new server can read or do, not just whether the assistant itself is on an approved list.

That is the useful direction here: the connection needs to be part of the policy, not just the application name.

Noon pricing and access

The public pricing page lists these plans:

Plan	Published price	Team size
Free	$0	Up to 10 users
Noon Pro	$7 per user per month	Up to 200 users
Enterprise	Custom	200 or more users

The page also advertises a 10% yearly discount. Free and Pro are subject to fair-usage limits, while Enterprise is priced per product with volume discounts. The Free plan lists email protection, phishing simulations, awareness training, device protection and AI governance. Pro adds features including deepfake voice simulations and same-day support.

Both Free and Pro currently lead to an early-access request. Enterprise leads to a demo. Published pricing therefore tells a buyer how the offer is structured; it does not establish that every visitor can start using every module immediately.

What to check before rolling it out

The first evaluation should be a coverage test. Can the product detect the specific assistants, extensions and remote connections the team actually uses? A broad claim about discovering AI tools is less useful than an accurate inventory of one real working environment.

Then test the response. Seeing an unapproved connection, warning about it and blocking it are different capabilities. A pilot should make clear which action happens in each supported environment, and whether an employee can understand and resolve the warning without opening a support ticket.

There is also a privacy tradeoff. A tool governing employee AI activity needs enough visibility to apply policy. Buyers should review what content is inspected, what is retained, which administrators can access it and how exceptions are handled. An attractive per-seat price does not answer those questions.

For small teams, the free tier could make that initial evaluation easier once access is granted. For larger organizations, the value depends on whether the unified product reduces fragmented controls without hiding gaps between them.

The meaningful test for Noon is whether an approved AI workflow stays usable while new connections become visible and reviewable. That is a more practical target than treating every new assistant as either completely trusted or completely banned.

I’m Paras. I work at the intersection of AI, engineering and distribution.

I help technical teams explain complex AI products through clear launch stories and founder-led content.

Book a 30-minute call.