# Identity & authority — AI Engineer Sydney 2026

> Whose authority does an agent act with, and where does it stop? Connect authentication, delegated permissions, fine-grained authorisation and audit trails with enforceable controls over tools, data and production changes.

Canonical page: https://webdirections.org/ai-engineer/topics/identity-authority/
Program status: The working timetable is public. Announced speakers and presentations are confirmed; some track assignments, session groupings and times may change as keynote, partner and specialist sessions are finalised. Morning keynote programmes and Industry sessions will be added as they are confirmed.

## Published talks

- [Identity for the AI Era: Securing Customers, Employees, and AI Agents](https://webdirections.org/ai-engineer/speakers/sean-spires/) — Sean Spires
  As AI agents take on more autonomous roles, the identity and access patterns built for humans and APIs no longer hold. This session draws on Auth0’s real-world experience securing agentic workflows where misconfigured permissions can have serious consequences. We’ll Cover - How Auth0 handles agent authentication and authorisation in real-world agentic…
- [Beyond System Prompts: Securing AI Agents with Fine-Grained Authorization](https://webdirections.org/ai-engineer/speakers/daizen-ikehara/) — Daizen Ikehara
  To make an AI agent useful, we often give it broad access to APIs and data. As systems move from simple retrieval to autonomous action, system-prompt instructions are not an adequate security boundary. This presentation examines Fine-Grained Authorization patterns for agents: how an agent can inherit a user's permissions, how identity should flow through…
- [MCP Isn't Enough: Why Agents Needed an Event Loop, Not Just an API](https://webdirections.org/ai-engineer/speakers/saul-caganoff/) — Saul Caganoff
  Many apps are adding MCP interfaces so agents can query their systems of record. It works, until the workflows we actually care about turn out to be "something happened, now respond", and polling wastes network, context and tokens. We bolted MCP onto our product, then added events. The industry is lagging: event publication is uneven, MCP's own event support…
- [From REST to Agentic: Trusted APIs in the age of AI](https://webdirections.org/ai-engineer/speakers/leigh-whiting/) — Leigh Whiting
  Enterprise platforms don't get to start from scratch. When AI agents need to act on behalf of users, they inherit the API surfaces those users already depend on — surfaces evolved over many years for human-driven workflows, not autonomous reasoning. This talk explores the practical journey of adapting established, trusted API layers to serve agentic…
- [Query before mutation: guardrail patterns for agents that touch production infrastructure](https://webdirections.org/ai-engineer/speakers/jeffrey-aven/) — Jeffrey Aven
  Giving an AI agent permission to change production infrastructure removes the human pause that made plan-and-apply safe. In this talk, Jeffrey Aven shows how to put that judgement back into the system itself. He introduces query before mutation: requiring an agent to establish the live state of an environment before it can write. From there, he develops a…
- [One harness for every coding agent: swap models and runtimes without rewriting](https://webdirections.org/ai-engineer/speakers/ananya-roy/) — Ananya Roy
  You want your team using the best coding agent for each job : Claude Code, Codex, Cursor, or one you built yourself , without rewiring your setup every time a better model ships, and without each agent running loose with its own credentials, spend, and idea of what's safe. Across the thousands of engineering teams we work with globally, the pattern is the…
- [Running AI Coding Agents at Scale: Adopting New Models, Controlling Cost, and Governing What Agents Do](https://webdirections.org/ai-engineer/speakers/anthony-ivan/) — Anthony Ivan
  Running coding agents across a company comes down to three questions: which models to offer, how to control cost, and what agents should be allowed to do. This talk gives four recommendations on those questions, drawn from running coding agents for 12,000 employees at Databricks. Offer new models early, and let data choose the default: every model reaches…
- [How to Count to One Hundred](https://webdirections.org/ai-engineer/speakers/hugo-oconnor/) — Hugo O'Connor
  Agents don't work well together in chat rooms: put a few in a channel and they talk past each other. Natural language and unrestricted JSON feed an open-ended pipeline of reasoning and tool calls with full Turing power. Deciding whether to accept a message based on what arbitrary computation will do is undecidable in general. Meanwhile, the communication…

Topic groupings connect talks across the programme. For session times, rooms and tracks, see the [schedule](https://webdirections.org/ai-engineer/schedule/). Programme details remain subject to change.

- [Explore the whole programme](https://webdirections.org/ai-engineer/program/)
- [Conference overview](https://webdirections.org/ai-engineer/index.md)