Guild.ai
San Francisco startup building a control plane that enforces identity, least-privilege access, and audit logging for autonomous AI agents inside enterprises.
Visit Website ↗ + Add to Compare Claim This CompanyOverview
Guild.ai markets itself as the first control plane purpose-built for AI agents: a governed runtime that enforces identity for each agent, applies least-privilege access to the tools and systems an agent can touch, and produces immutable audit logs of every agent action. Its Managed Agent Center and Agent Hub let organizations version, publish, and reuse agents internally rather than letting them sprawl ungoverned across teams.
The company was co-founded by CEO James Everingham, formerly a VP of engineering at Meta, and launched publicly in 2025. It raised a combined $44 million across seed and Series A rounds, announced in early 2026, with investors including GV (Google Ventures), Khosla Ventures, Acrew Capital, and NFX.
Guild’s thesis is that as enterprises deploy AI agents with real system access — to code repositories, ticketing systems, and internal data — those agents need the same identity, access-control, and audit rigor already applied to human employees, enforced through OAuth-based integrations with tools like GitHub and Jira.
Innovation Matrix Assessment
Shipped a novel, well-articulated 'control plane' framing for agent governance (identity enforcement, least privilege, audit logging) quickly after launch.
Directly addresses an emerging CISO pain point: AI agents operating with excessive, ungoverned privilege across internal systems.
A $44M combined seed/Series A from top-tier investors (GV, Khosla, Acrew, NFX) is a strong signal for a company only months past public launch.
AI agent identity and governance is a genuinely new sub-category rather than a repackaged point tool, though it is being contested by several similarly well-funded entrants.
The company is too new for public customer case studies or independent validation; all efficacy claims are vendor-stated.
Governance of autonomous AI agents is very likely to become more, not less, important to CISOs over the next three to five years.
Why CISOs Should Care
Gives security teams a way to govern AI agents the way they already govern human identities, enforcing least privilege and producing an audit trail before agents are granted broad system access.
What Makes It Different
Treats AI agents as first-class identities requiring their own dedicated control plane, rather than retrofitting existing human-IAM tooling onto agentic workflows.
The Matrix Verdict
60/100 — INCREMENTAL INNOVATOR
Incremental Innovator. Strong investor backing and a timely, well-articulated thesis, but the company is brand new with no public customer validation yet, in a sub-category several well-funded rivals are also contesting.
Editorial Note: Claims vs. Verified Findings
Founding year, headquarters, and funding breakdown conflict across aggregator sources (some list a much earlier founding tied to an unrelated open-source project); the $44M raise and investor list are corroborated by multiple press outlets, but no independent customer evidence exists yet.
Sources
Alternatives to Guild.ai
Adaptive Security
AI-driven platform that simulates deepfake, voice, and multichannel social-engineering attacks to train and test organizations against next-generation phishing.
Quilr
CISO ReviewedEarly-stage agentic AI security startup building a 'Service-as-Software' platform to guard against human-related breaches and secure AI agent…
Zenity
Governance and security platform for AI agents and low-code/no-code development, securing agent identity, permissions and behavior across the…
Tenzai
An agentic AI penetration testing startup building autonomous 'AI hackers' to find and validate exploitable vulnerabilities at a…
Gray Swan AI
Carnegie Mellon-linked startup building automated red-teaming and runtime guardrails that frontier AI labs use to stress-test their models…
Alice (formerly ActiveFence)
Israeli AI security company (rebranded from ActiveFence in January 2026) offering a lifecycle platform to test, guard, and…