Transparency
AI Use Disclosure
We build AI systems for a living, so we should be clear about how we use AI ourselves.
Last updated 19 November 2026
In our services
- We use AI models from established providers, through their APIs, inside the agents and automations we build for clients.
- Every agent has a defined autonomy level, a named human owner and documented escalation rules.
- Consequential actions require human approval unless the client has agreed, in writing, a higher autonomy level supported by evaluation results.
- We tell clients which AI providers are involved in their workflows and where data is processed.
In our research and content
- Our research and website copy are drafted with AI assistance.
- Third-party figures are checked against the original publisher before publication, and every source is linked.
- Forecasts are labelled as forecasts; our own frameworks are labelled as recommendations; examples are labelled as hypothetical.
On this website
This website does not use an AI chatbot or make automated decisions about visitors. If that changes, we will say so here.
Questions
Email hello@mirror.lk or, for research, research@mirror.lk.