Gardant × Teton Case Study · Noblesville, Indiana
Gardant at Noblesville
How anonymized computer vision reshaped fall safety at Gardant’s Noblesville community — a 69% reduction in fall rate across assisted living and memory care.
The program
Continuous visibility, built into the rhythm of care.
Gardant partnered with Teton to bring anonymized computer-vision observation to its Noblesville community, which brings assisted living and memory care together on one campus. Ceiling-mounted sensors process resident movement, position, and behavior on-device — no identifiable imagery ever leaves the room — so dignity is preserved while every covered space is observed around the clock.
The platform surfaces real-time alerts to staff devices for falls and other risk events, builds behavioral analytics on sleep, mobility, and activity, and includes Samwise, an AI chat-based assistant staff can query in natural language.
Installation rolled out across the community in early October 2025. After a stabilization period, the program ran a disciplined 34-day baseline before entering active use, so every result could be measured against the same residents and the same staff.
Operator
Gardant Senior Living
Community
Heritage Woods / White Oaks of Noblesville
Setting
Assisted living & memory care
Install
October 2025
Measurement
34-day baseline vs active use
On-device vision
Processed locally; imagery never leaves the room.
Real-time alerts
Falls and risk events routed to staff devices.
Behavioral analytics
Sleep, mobility, and activity patterns over time.
Samwise
Natural-language assistant, proactive insights.
How the pilot was run
A defensible before-and-after, measured against the same community.
Oct 2025
Install & onboarding
Sensors, network, and staff training across the community.
Dec 2025 – Jan 2026
Baseline window
A 34-day pre-period with no intervention, the control.
Feb – Apr 2026
Active use
Alerts, digital rounding, and analytics fully live.
Outcomes at a glance
Measurable gains across every dimension the system tracks.
69%
Fewer falls per resident — the fall rate fell across the pilot, even as census grew.
97%
Faster reaction time — a resident on the floor is now reached in under 3 minutes on average.
92%
Less time on the floor — lay time after a fall collapsed to under 4 minutes.
80%
Reached under five minutes — most falls are now answered within five minutes.
100%
monitoring coverage maintained throughout the pilot, across assisted living and memory care
~28
safety-relevant notifications per resident per day, far more for higher-risk residents
820
digital spot checks completed (85% rate): timestamped, auditable touchpoints
Safety outcomes
Fewer falls and faster response, even as the census grew.
The resident population grew over the pilot, yet the fall rate per resident dropped 69%. Because the census rose rather than shrank, that decline is a genuine safety gain, not an artifact of fewer people to fall.
69%
fewer falls per resident, even as census grew
13%
preventable fall rate: staff intervening on risk before incidents occur
40%
share of all falls from repeat fallers, the hardest concentration to move
18%
residents falling more than once during the period
Reading the repeat-faller numbers
A small number of residents typically generate a disproportionate share of incidents. With repeat-faller concentration down to 40% and preventable falls down to 13%, the platform is helping staff intervene earlier on at-risk residents, not just respond faster after an event. The residual reflects the baseline risk inherent to this acuity level.
What changes after a fall
“When a caregiver reaches a fallen resident in minutes, getting up safely becomes a decision, not a gamble.”
40%
self-recovery rate: fewer residents left to get up unaided
92%
less time on the floor after a fall
Because lay time collapsed, residents are now far more often assisted up by staff than left to recover alone, with self-recovery now accounting for 40% of falls. Assisted recovery is materially safer: a resident getting up unaided risks a secondary fall or a worsened injury, while a caregiver arriving promptly can stabilize them, assess for injury before any movement, and catch problems a resident might not report.
Prompt arrival also sharpens incident review. Staff observe the undisturbed scene and pair their account with the anonymized clip from the platform. That visual context informs ER decisions too: footage of a controlled lowering or a non-injurious fall lets staff make a more confident call and avoid an unnecessary hospital transfer.
Beyond falls
Healthier residents, and a more deliberate way of working.
Resident health & wellbeing
51 → 55
Walking-speed score — an early signal of healthier mobility.
49 → 52
Stationarity score — more balanced movement through the day.
11.5 → 11.8
Hours of sleep / night — longer, steadier rest across residents.
Staff workflow
12.7 → 9.7
visits per resident per day
3.1 → 3.4
minutes per visit
Visits dropped, but each one grew longer and more purposeful: attention directed where it matters most, rather than spread evenly across the floor.
What it took
Two things had to go right.
01
Embedding into the culture, not on top of it.
Getting training right took multiple on-site visits from Teton’s Customer Success Managers: understanding existing workflows and finding the right way to fold Teton into the day-to-day rather than layering it on top. Sustained adoption depended on staff seeing the platform as a tool that enhanced their work, which took deliberate iteration on workflows, alert configuration, and training cadence.
02
A defensible baseline in a live community.
Establishing credible numbers meant running a disciplined 34-day pre-period before activation, which required operator commitment to deferred value in service of clean measurement. Most deployments skip this in favor of reaching outcomes faster, but the structured baseline is exactly what makes the before-and-after credible: the same residents and the same staff, measured against themselves under identical operating conditions.
Why it matters
Breadth is the difference.
Most monitoring systems in long-term care focus narrowly on a single signal — a fall, a wandering event, a pull-cord — and operate reactively. The Teton platform at Heritage Woods and White Oaks continuously observes movement, position, sleep, mobility, activity intensity, and state-level risk, then routes the right signal to the right staff member at the right time.
The repeat-faller numbers show why that matters. Dropping repeat-faller share to 40%, alongside the 69% overall reduction and the gains in sleep, mobility, and activity, indicates the monitoring layer is enabling staff to intervene earlier, not just respond faster. State detection for risky activity, paired with the behavioral data layer, is what makes that preventive work possible: a meaningful step beyond traditional reactive monitoring.
What the platform observes
- Movement & position
- Sleep & rest
- Mobility & gait
- Activity intensity
- State-level risk
Prevention belongs at the point of care.
If you’d like to see what Teton would reveal inside your community, we can walk you through what the first ninety days typically look like.
Book a demo