Responsible AI at Say4
The SafetyUX AI Engagement Assistant helps people connect evidence, prepare a response and keep action moving. Its role is advisory: accountable people remain in control.
See the boundaries ↓A clear role
The assistant can help
The assistant does not
Our product principles
These principles guide product design and implementation. Exact technical, privacy and supplier controls are documented for the agreed use case.
Use information the organisation has approved for a defined and relevant purpose.
Make recommendations available to accountable people to accept, change or reject.
Help reviewers see the goal, sources and themes informing a suggestion.
Use AI where it supports preparation and insight—not where it would displace safety-critical judgement.
Before using an output
Every meaningful suggestion should be reviewed in the context of the organisation, the quality of the evidence and the people who may be affected.
Read the full responsible AI guideRelevantDoes it address the agreed question?
GroundedWhich approved information supports it?
ProportionateIs the action justified by the evidence?
FairWhat context or perspective might be missing?
Safe to useDoes this need competent operational or legal review?
OwnedWho accepts, changes or rejects it?
Data and implementation
Define the task, the intended benefit and the decisions the assistant is not authorised to make.
Agree sources, access, retention, deletion and whether individual-level information is needed at all.
Set review ownership, escalation points, feedback routes and a cadence for evaluating performance.
Say4 will not use unverified marketing language to imply certifications, regulatory approval or technical controls that have not been confirmed. Product and data-processing details will be documented for each implementation.
Standards we follow closely
Discuss the assistant