These principles describe our approach to building and evaluating AI. Actual functionality and safeguards must be confirmed for each product and deployment.
Human judgment matters
AI can organize information, identify patterns and prepare work. Consequential decisions need an accountable person. Our product direction prioritizes human review where a decision affects a customer, an agent’s standing, compliance or another material interest.
Understand the limits
AI outputs can be incomplete, inaccurate or inappropriate for a particular context. Users should verify material facts and use qualified professional judgment before relying on them.
A recommendation, summary or score should not be treated as a carrier decision, legal conclusion, licensed insurance recommendation or guarantee of an outcome.
Use data deliberately
Appropriate data access, permitted purposes and clear handling rules should be established before a deployment. Avoid supplying information that a team is not authorized to process.
Whether inputs, outputs or feedback are used to train or improve a model depends on the specific product, provider and agreement. Confirm those terms before uploading customer information; this page makes no blanket training or data-residency promise.
Design for accountability
Our development approach values traceable actions, understandable context, human review and a way to raise concerns. The controls appropriate for a workflow should be agreed, implemented and evaluated before expanding automation.
Responsible use
AI should not be used for deception, unlawful discrimination, unauthorized surveillance, impersonation, fraud or to bypass privacy, consent, licensing or security requirements.
Users remain responsible for the permissions, professional duties and legal requirements that apply to their activities. Product agreements may set additional restrictions.
Keep the conversation open
Raise concerns about a proposed use, an output or a workflow with the relevant product team. For initial discussions, use Get Started. Avoid including sensitive customer information in a general enquiry.
For website privacy questions or security concerns, see our Privacy Policy and Security information.
Clear information. Considered choices.
Back to Trust at NeoStruX ↑