We will cover the foundations that made the shift possible. First, an AI policy designed to enable experimentation rather than block it. Second, shared organisational memory: Memra, an internal platform that combines retrieval and a knowledge graph so people and agents can draw on context across conversations, projects, and accounts. Through a live demo, we will show how agents now take part in everyday work, from shaping proposals and spinning up working Drupal proofs of concept to automating operational workflows, with humans owning the decisions that matter.
We will also be honest about what went wrong: confidently wrong answers, data bleeding across accounts, security gaps in our own bots, and the cultural resistance no tool can fix. Finally, we will look at how roles and hiring change when AI rewrites which skills are rare.
Attendees will leave with a practical playbook for moving their own teams from scattered AI usage to an AI-native way of working.
Learning outcomes
By the end of this session, attendees will be able to:
- Distinguish AI-assisted from AI-native teams, and explain why process and culture matter more than tooling
- Design guardrails that accelerate adoption, including an AI policy that enables rather than restricts
- Understand how shared organisational memory (retrieval plus a knowledge graph) makes AI useful beyond individual prompts
- Apply human-in-the-loop patterns that keep agents safe, accountable, and trustworthy
- Anticipate common failure modes in internal AI systems and how to test for them
- Start the shift in their own organisation with a step-by-step playbook