Agentic Fabriq
Okta for Agents.
Winter 2026ActiveB2BAIOpsDevSecOpsSaaSAPIInfrastructureSunnyvale, CA, USA
Agentic Fabriq is the single control plane for data permissioning throughout your entire organization across your AI and employees. Through Fabriq, any employee can tap into AI/agents with exactly the data and tools they're cleared to use by an admin. Fabriq can be used internally with employees or externally for customers while handling the auth + integrations + permissions securely for you.
Verdict
High Signal
Market Opportunity
AI agent security and identity/permissioning is a real and rapidly growing B2B enterprise problem. ICP is clear: IT/security teams at companies deploying AI agents. The 'Okta for agents' framing maps to a proven $20B+ market segment, and agentic AI proliferation makes this urgent. Monetization path is obvious via SaaS per-seat or enterprise contracts.
Medium Signal
Founder Signal
Matthew Xu dropped out of MIT at 19 — young with minimal LinkedIn data available. Paulina Xu is AI + Physics at MIT, still a student. No prior work experience, exits, or shipped products documented in the provided data. Strong academic pedigree but zero work history to evaluate — this is a student-founder team at best.
Low Signal
Competition
No competitor data was returned, but the space is heating fast. Direct competitors include Permit.io, Indent, and emerging agentic-specific players. Okta itself, along with Microsoft Entra, could extend into this space natively. AWS, Google, and Anthropic are all building agent management layers. The 'Okta for agents' positioning is already being used by multiple startups simultaneously, suggesting a crowded framing if not yet a crowded product.
Medium Signal
Product
Website has a pricing page, detailed feature descriptions (per-user identity binding, real-time guardrails, tool permissioning), TypeScript & Python SDKs mentioned, and 'sign up for access' CTA. No named customer logos, revenue figures, or live demo visible — just a demo booking flow and early access signup. The 88% stat about agent incidents appears to lack a cited source.
OverallC Tier
The market thesis is sound — AI agent permissioning and identity is a genuine enterprise pain point and the timing is right. However, the founding team is two MIT students with no documented work experience, no prior company, and no LinkedIn data to evaluate. There are zero customer logos, revenue signals, or proof of product-market fit beyond a demo-booking site. The competitive risk is severe: Okta, Microsoft, AWS, and a dozen well-funded startups are converging on this exact problem. Without a technical moat, proprietary data, or early enterprise traction, this looks like a smart idea from smart students that will be outcompeted by teams with deeper enterprise security credibility.
