Emergence: the factory infographic
Agentic infrastructure for enterprises: verified, governed agents. Speaking on 'When AI agents go rogue'.
What the company does
In normal terms
Builds agentic infrastructure for enterprise: verified, governed AI agents that plan, reason and act across complex systems.
Emergence builds infrastructure for enterprise agents that plan and act across complex systems, with verification and governance.
In lay terms
Emergence builds AI workers that can do whole jobs across a company's systems, with checks built in.
The risks it faces, especially from AI agents
In normal terms
- Agents that plan. Planning agents choose their own steps; the path they take may cross systems nobody expected.
- Going rogue. An agent acting within its reach but outside its mandate is the 'rogue' case their talk names.
- Proving 'verified'. Enterprises need to see what verified means for each agent.
In lay terms
- Agents that plan. Ask it to book a meeting and it decides to reorganise the calendar.
- Going rogue. It did nothing it couldn't do, just things nobody asked.
- Proving 'verified'. 'It's been checked': by whom, against what?
How RiskMandate helps
- Reachwhat it can actually access or affectEvery system an Emergence agent can plan across.
- Mandatewhat the business authorisedThe goal the business set.
- Gapthe difference: the riskPlans that leave the mandate: the rogue zone.
- Barrierswhat really stands in the wayIts verification and governance, as evidence.
- Provenancewho owns it, who reviewed it, which versionOwner and version of each agent.
In normal terms
Their talk asks what happens when agents go rogue; RiskMandate's answer is the Gap: know the reach, define the mandate, control the difference. A story partner and a possible joint example.
In lay terms
An agent goes rogue when it does what it can, not what it was asked. RiskMandate measures that difference.
A first line for the conversation: Your talk is 'When AI agents go rogue'. Rogue is the gap between reach and mandate. We measure it.
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): 'When AI agents go rogue', 12 Nov, AI Summit, Stage 13, names Emergence as the speaker's organisation
- websummit.com, read 2026-11-01: LIS26 attendee page: Satya Nitta, Co-founder, Executive Chairman & Chief Scientist, Emergence; country and industry as listed by Web Summit; public bio
- www.emergence.ai, read 2026-11-01: what the company does, from its own site (title and description)
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.