78%
Check-in completion rate
Illustrative pilot figure
Programme view
Admin / programme dashboard
Illustrative, anonymised aggregates for programme design. All numbers are invented for this prototype and describe no real person.
412
Guided demos completed
Aggregate count, no identifiers
36
Help requests raised
All routed to a human officer
4
False escalations under review
Reviewed manually each week
Accessibility settings used
Percentage of sessions in which each setting was active. Used only to improve the design.
- Larger text enabled62%
- Slow speech enabled45%
- Captions/transcript opened71%
- Reduced motion enabled18%
- Non-English interface selected39%
False-escalation review
Every escalation that turned out to be unnecessary is reviewed by a human to improve the rules — never to score or rank a participant.
- Case 1: reminder repeated during a hospital stay — quiet-hours rule updated.
- Case 2: missed check-in while travelling — added a “away” setting.
- Case 3: contact notified twice — de-duplication fix queued.
- Case 4: scam-check flow exited early — wording simplified.
Out of scope by design
No emotion recognition, no medical diagnosis, no social scoring or ranking of participants, no automated eligibility or benefit decisions, and no re-identification of individuals from these aggregates.