Agents in Production: The Auth Stack for AI — SF Tech Week 2026
Agents in Production: The Auth Stack for AI — SF Tech Week 2026
Thursday, October 8, 2026
San Francisco
AI agents are moving quickly from demos into production. But once an agent can access company data, call APIs, use business applications, and take actions on behalf of employees or customers, a new infrastructure problem emerges:
How do you control what that agent is allowed to do?
Join Agentic Fabriq and Open Future Forum during SF Tech Week 2026 for a technical discussion on the emerging identity, authentication, authorization, and governance stack for production AI agents.
This event is designed for founders, CTOs, engineering leaders, security leaders, AI infrastructure teams, and developers building or deploying agents in real-world environments.
What We’ll Discuss
Moving an agent from prototype to production introduces challenges that traditional application security was not designed to solve.
We’ll explore questions including:
How should AI agents authenticate with applications, APIs, and enterprise systems?
What identity should an autonomous or semi-autonomous agent have?
How should agents act securely on behalf of human users?
What permissions should an agent receive, and for how long?
How do you implement least-privilege access for agents?
How should credentials, secrets, tokens, and delegated access be managed?
What happens when agents dynamically invoke new tools?
How do companies monitor and audit what agents actually do?
Where should approval workflows and human-in-the-loop controls sit?
What does the emerging auth stack for AI agents look like?
From Demo to Production
Building an impressive agent demo is increasingly straightforward.
Deploying that same agent inside an enterprise is different.
Production agents may need access to systems such as Salesforce, Slack, Google Workspace, GitHub, databases, financial systems, internal APIs, and hundreds of other tools.
Each connection creates questions around identity, permissions, authorization, security, and governance.
As agents become capable of taking increasingly consequential actions, these controls become part of the core AI infrastructure stack rather than an afterthought.
This event will focus on the practical architecture required to make that transition safely.
Who Should Attend
This gathering is particularly relevant for:
AI and agentic startup founders
CTOs and technical founders
VP Engineering and engineering leaders
AI infrastructure teams
Security and identity leaders
Developers building production agents
Platform and enterprise architects
Investors focused on AI infrastructure
Format
Expect a focused technical conversation followed by drinks and networking with other builders working on AI agents, infrastructure, security, and enterprise deployment.
No generic AI predictions. No product-pitch marathon.
The goal is to bring together people solving the difficult infrastructure problems that appear once agents start doing real work.
Presented by Agentic Fabriq
Agentic Fabriq is building identity, access, and governance infrastructure for AI agents, helping organizations control what agents can access and what actions they are authorized to take across enterprise systems.
In Partnership with Open Future Forum
Open Future Forum is a global executive community founded in Silicon Valley. It brings together founders, executives, investors, and technology leaders through curated events, private forums, peer groups, and original research.
This is an independent community event taking place during SF Tech Week 2026. It is not organized or endorsed by Andreessen Horowitz.