Research
Care quality should be measurable, comparable and improvable.
What elderly care lacks is not commitment but a method for turning commitment into standards. Our research addresses one subject: how to make care quality countable, comparable and convertible into improvement.
- 6
- questions for the next five to ten years.
- 1
- published line on standards and definitions.
- 0
- figures without a source.
None of the six has a settled answer, and some do not yet have a settled form of asking. They are published because an agenda is itself a commitment.
- One
- The computability of care: the 90 metrics are a starting point; whether competent care can be turned into a measurable quantity.
- Two
- Learning without moving data: on-device training and model evolution under data-residency rules.
- Three
- Knowing without watching: non-contact sensing and millimetre-wave radar that flag risk without surveilling residents.
- Four
- Dividing responsibility before a machine acts first: human-machine boundaries must be testable, not merely stated.
- Five
- Turning the cost of one server into a monthly line: TCO must keep falling, because infrastructure is only infrastructure when it is affordable.
- Six
- Whether a machine can enter the corridor by 2035: the feasibility of embodied care in night patrols and transfers (planned).
Standards
Establish the standards and definitions before discussing intelligence.
Definitions come from published standards rather than our own formulation. The 90 metrics map item by item to features and ledgers; every cited document carries its full title and number; where a source is missing, it is marked as pending.
- Policy
- Hu Min Bian Han [2026] No. 74, 90 items, mapped one by one.
- Standards
- Dahua Welfare Home was a core drafter of the 2022 smart eldercare group standard.
- Data
- Every industry figure carries institution and year; forecasts are labelled as forecasts.
- Gaps
- Sources that cannot be found are marked as pending, never replaced by approximations.
We publish only verifiable milestones: start in 2025, validation and launch in 2026, commercial rollout in 2028, multi-scenario AI and the data flywheel in 2030, embodied care in 2035. All are plans, not achieved results.
- 2025
- Programme start; AI-native OS architecture established.
- 2026
- Validation and launch: digital workers live, Ark deployed, high-availability testing.
- 2028
- Commercial rollout, beginning in Beijing, Shanghai, Guangzhou and Shenzhen.
- 2030
- Multi-scenario AI and the data flywheel.
- 2035
- Embodied robotic care and full-element intelligent scheduling.
Published material
Only verifiable material is published.
The technical whitepaper covers five parts: architecture, model, posts, metrics and deployment. The complete version is provided during demo and POC. The former standalone whitepaper page has been folded into this section.
Book a demo
Title
Thirty minutes to observe nine digital workers operating in a real care setting. An advisor will call to confirm the time and focus.