Linux Foundation (Agentic AI Foundation): the factory infographic
A neutral home for open standards in agentic AI.
What the company does
In normal terms
The Linux Foundation hosts open-source projects; the speaker is Executive Director of its Agentic AI Foundation (AAIF), a neutral foundation for open standards in agentic AI.
The AAIF is a neutral foundation for open standards and projects in agentic AI, under the Linux Foundation.
In lay terms
The Linux Foundation is where companies agree on shared rules, like agreeing on the size of plugs so everything fits.
The risks it faces, especially from AI agents
In normal terms
- No common policy format. Each vendor describes agent permissions differently, so policies cannot travel between tools.
- Standards that ignore the business. Technical standards may cover identity and protocols but not what the business authorised.
- Slow consensus. Agents spread faster than standards are agreed.
In lay terms
- No common policy format. Every country with its own plug.
- Standards that ignore the business. The plug fits, but nobody says what you're allowed to plug in.
- Slow consensus. The appliances arrive before the plug rules do.
How RiskMandate helps
- Reachwhat it can actually access or affectA standard way to describe what an agent can access.
- Mandatewhat the business authorisedA standard way to state what it may do.
- Gapthe difference: the riskComputed the same way everywhere.
- Barrierswhat really stands in the wayEvidence fields any tool can fill.
- Provenancewho owns it, who reviewed it, which versionSigned owner, reviewer and version.
In normal terms
RiskMandate's Agent Behaviour Policy could become an open spec: Reach, Mandate, Gap, Barriers and Provenance as a common format. Ask the AAIF how a policy format becomes a standard.
In lay terms
RiskMandate's one-page agent record could become the standard plug for AI safety.
A first line for the conversation: Agents need a common way to say what they may do. Could a behaviour policy format become an open spec?
Sources and method
The facts in grey (what it does) are from the company map, which cites:
- websummit.com, read 2026-11-01: LIS26 session page (conferenceSlug lis26): 'Why AI agents stall at the production line', 11 Nov, AI Summit, Stage 13, names Linux Foundation as the speaker's organisation
- websummit.com, read 2026-11-01: LIS26 attendee page: Mazin Gilbert, Executive Director, Linux Foundation; country and industry as listed by Web Summit; public bio
- aaif.io, read 2026-11-01: what the company does, from its own site (title and description)
- www.linuxfoundation.org, read 2026-11-01: what the Linux Foundation is, from its own site
The risks and how RiskMandate helps are our analysis, written by dev@web-summit.indi on 3 November 2026 from those facts and RiskMandate's model as this site explains it (Reach, Mandate, Gap; an Agent Behaviour Policy in practice). They are not the company's own statements and name no incident. A person reviews each before it is used.