Agentic Fabriq

What do they actually do

Agentic Fabriq provides a developer-installed permissioning and audit layer for AI agents. Teams add a TypeScript or Python SDK, register their agents and tool connections, and set per‑role policies so agents act only with the permissions of the user who invoked them. The platform enforces least‑privilege scopes on every call and records searchable, signed logs of agent actions (site).

Admins get SSO integration, role‑based access controls, instant revocation, and centralized auditability. The company is very early and is currently booking demos and running private pilots rather than broad self‑serve availability (site, pricing, YC profile/launch).

Who are their target customer(s)

• IT/Security managers at mid‑to‑large companies: Need to control who can run agents and exactly what those agents can access; want per‑user credentials (no shared creds), centralized policies, SSO, instant revocation, and least‑privilege enforcement.
• Engineering teams building agent‑powered apps: Don’t want to rebuild OAuth and per‑tool permissioning for every integration; need SDKs/connectors so agents run with the caller’s scopes without custom auth plumbing.
• Platform/DevOps teams running internal automation: Must trace, debug, and roll back agent actions; need searchable, signed logs and role‑based policy controls to investigate incidents and change access immediately.
• Compliance and audit teams in regulated industries: Require provable, per‑user trails for every agent action; need auditable evidence tied to users and policies for reviews and regulators.
• Early‑stage product/AI teams running pilots: Need to move fast without insecure prototypes; want a plug‑in permission layer so pilots use per‑user, scoped access while validating value with users.

How would they acquire their first 10, 50, and 100 customers

• First 10: Founder‑led outreach to YC/network and engineering leads running agent pilots; offer a 4–8 week pilot where Fabriq ships the SDK hookup and one critical connector, booked via the demo/pilot CTA (site, YC profile).
• First 50: Package a repeatable pilot (checklist, a small set of high‑value connectors, sales‑engineer playbook) and source candidates via developer/security communities and events while hiring the first sales engineer to run onboardings at scale (site).
• First 100: Layer in enterprise procurement: deepen SSO/IdP integrations, add a partner/channel program, and introduce a simple self‑serve tier for smaller teams so sales focuses on larger pilots and contracts (site, YC launch).

What is the rough total addressable market

Top-down context:

Agent permissioning sits within Identity & Access Management and Privileged Access Management. Recent estimates place IAM in the low‑to‑mid tens of billions of dollars and PAM in the low single‑digit billions, both growing as enterprises secure more SaaS and automation (IAM, PAM, security spend context).

Bottom-up calculation:

Using ~$18B as a blended IAM+PAM reference, if agent‑specific controls capture 5–10% as automation adoption rises, the SAM is ~$0.9–$1.8B. If Fabriq executes well, capturing ~0.5–2% of that SAM over 3–5 years implies roughly $4–$36M ARR, consistent with a focused, early enterprise product (product focus, IAM, PAM).

Assumptions:

• Enterprise prices for pilots/early deployments are in the tens to hundreds of thousands per year; no public pricing yet (pricing).
• Adoption depends on shipping enterprise features (connectors, SSO/IdP support, tenant isolation, signed/searchable logs) that reduce integration risk (site).
• Large identity/cloud vendors may bundle adjacent capabilities; Fabriq must win on developer ergonomics and depth of connectors.

Who are some of their notable competitors

• Keycard: Positions a unified identity and policy layer for AI agents with developer SDKs and end‑to‑end auditing—directly overlapping Fabriq’s core use case (site, background).
• Zenity: Focuses on discovering agent activity, step‑level observability, and automated detection/response for risky agent behavior—competes more on security operations than SDK integration (site).
• OASIS Security: Offers non‑human identity management and automated provisioning/rotation for service and agent accounts across clouds—overlaps on lifecycle and policy for machine/agent identities (site).
• Credo AI: AI governance and compliance platform providing policy management, risk assessment, and audit evidence—competes for compliance buyers rather than per‑call SDK enforcement (site).
• Okta: Incumbent identity provider; now supports non‑human/AI agent identities, SSO, and privileged access features—many enterprises may extend existing Okta deployments instead of adding a specialist layer (Okta).