Senior Living · Operations & Admin
You are already giving the care. You are not always billing for it.
A resident who needed one aide for transfers in February needs two by June. The aides know it. The care plan does not say it. So you staff at the higher level, bill at the lower one, and the gap runs until somebody remembers to reassess — often at the annual review, four months later, across every building at once.
That is the largest quiet number in most assisted living operations, and it lives in the same place as the 5:12am call-out, the survey binder assembled in three panicked days, and the move-in that slipped ten days waiting on a physician's form. None of it is clinical work. All of it is reading documentation, noticing a mismatch, and chasing someone.
Four places the margin leaks out of a building
Acuity climbs and the service plan stays where it was
Your aides document the change honestly — two-person transfer, new incontinence care, a med pass that now takes fifteen minutes instead of five. That documentation sits in the record and nothing downstream reads it. The resident's assessed level of care, and the rate attached to it, keeps describing a person who no longer exists.
The 5:12am call-out becomes an agency shift by 6:30
Your scheduler starts down a list, gets three voicemails and a no, and calls the agency because the shift starts in ninety minutes. Nobody had time to check who is under their overtime threshold, who picked up last Saturday and is owed the courtesy of not being asked again, or which med tech is already in the building on a double.
Survey readiness is a three-day scramble, not a state
A surveyor walks in and somebody stops doing their job to assemble service plans, med error logs, incident reports, fire drill records, and staff training files into a binder that was supposed to be current all along. You usually pass. The cost is not the tag — it is the three days and the fact that nobody could have told you on Monday whether you would.
A move-in slips ten days on one outstanding form
The family is ready. The apartment is ready. The physician's report has not come back from an office that faxes, and nobody owns chasing it. Ten days of an empty apartment on a unit you have already marketed, staffed, and budgeted — and a family whose enthusiasm is quietly decaying the whole time.
Reading the record you already keep
Everything on this page starts from the same observation: your buildings already generate the evidence. Aides document. Nurses chart. Incident reports get filed. Schedules get built. The failure is not that the information does not exist — it is that no human has time to read all of it, across every resident and every shift, looking for the mismatch.
So we build agents that read. One watches care documentation against assessed levels and flags residents whose recorded needs have drifted past what their service plan says, with the specific notes attached so your nurse can reassess in ten minutes instead of starting from a blank form. It never assesses anybody and never touches the clinical record — it points, a licensed person decides.
The same shape covers the rest. A scheduling agent works your internal roster in the right order — qualification, overtime exposure, who was asked last — before agency is ever called, and logs why each person was passed over. A readiness agent keeps a running survey posture per community, so the answer to 'are we ready' is a number on a Monday rather than a scramble on a Thursday. A move-in agent tracks every outstanding document per pending admission and chases the physician's office on a cadence a human would not sustain.
What we build toward
- 3–5 daysOff the move-in cycle
- Fewer days between a family saying yes and a resident actually moving in, because the outstanding documents get chased the day they age rather than the day someone notices. On an apartment you have already staffed and budgeted, those days are pure margin.
- Every shiftInternal roster exhausted before agency
- Agency becomes the genuine last resort rather than the fastest option at 5:30am, with a logged reason for every internal person who was passed over — which also happens to be the record you want the next time someone asks why the labor line moved.
- ContinuousSurvey readiness, as a number
- A standing posture per community instead of a three-day assembly project. You find out that Building 4's service plans have drifted in the week it happens, not in the week the surveyor arrives.
Targets, not results. Care-level drift is the one worth measuring first because it is usually the biggest and it is the easiest to verify — we can size it from a sample of your own records before you commit to anything.
What actually gets built
Care-level drift detection
Reads aide documentation, shift notes, and incident reports against each resident's assessed level of care, and flags the mismatches with the supporting entries attached. Your nurse reassesses; the agent never assesses, never writes to the chart, and never proposes a rate.
Call-out and open-shift handling
Works your internal roster in a defensible order — qualification for the position, overtime exposure, rest since last shift, how recently they were asked — by text, and escalates to agency only once the internal list is genuinely exhausted. Every decision is logged.
Continuous survey posture
Checks what a surveyor would actually ask for — current service plans, med error logs, incident follow-up, drill records, training files — on a schedule, and reports a readiness state per community rather than producing a binder on demand.
Move-in document chasing
Tracks every outstanding item per pending admission — physician's report, TB clearance, the state assessment form, financial paperwork — and chases the family and the physician's office on a cadence, then tells the executive director which move-ins are at risk this week and exactly why.
Family communication drafting
The post-incident call, the change-in-condition note, the quarterly care conference summary. Drafted from the record in your community's voice, sent by the person whose name is on it.
Modeled, not claimed
What care-level drift costs across three communities
Three buildings, roughly 120 residents combined, a tiered level-of-care schedule with meaningful steps between tiers.
Assume that at any given time some share of residents are receiving documented care above their assessed level, and that the drift persists for a few months before a reassessment catches it. You are paying for that care in labor either way — the only question is whether the rate reflects it.
Put a modest share and a modest number of months against your own tier steps and 120 residents, and the annual figure typically lands well past the cost of building the detection. It is also the rare project where the finance case does not depend on anybody working faster, because nothing about the care changes — only whether the assessment kept up with it.
- 120
- residents nobody can read weekly by hand
- Months
- typical lag before drift is caught
- Sample
- we size it from your own records first
Arithmetic, not a case study, and deliberately left without a percentage — the honest version of this number can only come from a sample of your records. We will run that sample before proposing anything, and if the drift is not there we will tell you.
What operators push back on
Our nurses assess residents. I am not putting software in the middle of a clinical judgment.
Neither are we, and the design says so: the agent reads documentation and raises a flag with the underlying notes attached. It does not assess, does not score acuity, does not propose a level, and does not write a word into the clinical record. What it replaces is the impossibility of one nurse re-reading every resident's documentation every week. The judgment stays exactly where your license says it has to be.
How does this hold up in a survey?
Better than the current state, because the audit trail is the point. Every flag, every reassessment it prompted, and every internal staffing decision it logged is timestamped and attributable to the human who acted on it. Surveyors are not troubled by a system that helps you notice things sooner; they are troubled by records that cannot explain themselves.
We run PointClickCare in two buildings and ECP in the third. That is the actual problem.
It is a problem and it is also normal, usually the residue of an acquisition nobody has had time to unwind. We read from both rather than making you consolidate first — the agent works off each system's API or export and normalizes on our side. You will get a straight answer during scoping about which of your systems is genuinely painful, and there is usually one.
Straight answers
Does this write anything into the EHR or the eMAR?
No, and that is a deliberate scoping decision rather than a technical limit. Everything it produces is a flag, a draft, or a report that a licensed human acts on inside your system of record. Write access to a clinical record is a category of risk that buys you very little here.
Will the scheduling piece respect our union rules or our internal seniority policy?
Yes — those rules are the ordering logic, not an exception to it. We encode your actual contract or policy during scoping, and the log for every open shift shows the order it worked and why each person was skipped, which is precisely the record you want if the order is ever challenged.
Does memory care work differently from assisted living here?
The care-level piece does. Memory care tends to have a different tier structure and behavioral documentation that reads nothing like ADL documentation, so it gets its own drift logic. The staffing, survey, and move-in pieces are largely the same across both.
How long before this is running in all our buildings?
One building live in about six weeks, and the additional communities are mostly a data-access exercise after that — two to three weeks each, longer if a building is on a different system. We deliberately start with one and let your team distrust it in public for a few weeks before it goes anywhere else.
Start with the drift number
The Pulse AI Operations Index scores your operation across intake, admin, and reporting — and for senior living it starts by sampling your own records to see whether care-level drift is real in your buildings. Twenty minutes.