Research Assistant
Literature review and technical summaries.
AI-native organisation
AI agents & automation.
An AI-native way of working, with AI agents handling routine tasks so human professionals can focus on engineering, judgment and safety.
Long-term target
Operating model
Humans and AI, working together.
Our long-term target is for roughly half of operational tasks to be supported by AI agents and automation, with the other half performed by human professionals.
This is an aspirational target, not a measured figure.
Potential AI agents
Nine roles we are exploring.
Software Development Assistant
Code drafting, review and testing support.
Data Analyst
Exploring and summarising datasets.
Project Management Assistant
Plans, schedules and status tracking.
Documentation Agent
Keeping technical documents current.
Marketing Agent
Drafting content for human review.
Customer Support Agent
First-line responses and routing.
Quality Assurance Assistant
Test planning and issue detection.
Recruitment Assistant
Scheduling and screening support.
Human accountability
People keep the final say.
AI agents assist; humans decide. Humans retain responsibility for:
- Critical decisions
- Engineering validation
- Financial approvals
- Safety
Guardrails
- Human review of agent output before it is relied on
- Access limited to what each agent needs
- Logs of agent actions for audit
- Careful handling of confidential data