Put AI on the RPM work that doesn't bill.
From January 2027, CMS proposes that billable monitoring minutes come from the billing practice's own employees, not a contracted vendor's. The chasing that fills a coordinator's day, the readings that stopped transmitting, the fourth reminder, the first pass over a reading that turns out fine, is exactly the work that rule reprices. AnyBio puts governed agents on it: the outreach and symptom check-ins run automatically, every message is checked before it sends, and your clinicians keep the calls that need a license. Built for the platforms and providers who run RPM.
Contracted staff gets repriced. The work doesn't go away.
Remote monitoring pays through two kinds of code: one for supplying the device and moving the data, one for your clinical staff's time with the patient. The chasing in between belongs to neither. Today it falls to a care coordinator or a contracted nurse. From 2027, the proposed rule means the contracted version stops counting toward the minutes a practice can bill.
The work still has to happen. The question is what does it: another hire the panel will outgrow, or software your own clinicians run over the top. AnyBio is that software. It chases the readings, sends the reminders, and collects the first round of symptoms when a reading lands out of range, so the nurses and coordinators you employ spend their time at the top of their license, on the patients who actually need them.
- +Chases the patients who stopped transmitting
- +Sends and escalates adherence reminders
- +Checks in when a reading lands out of range, and collects symptoms
- +Drafts the note for your clinician to review
None of it needs a license. All of it is why panels stop growing.
- +Interactive communication with the patient
- +Clinical judgment on what a reading means
- +Any change to the care plan
- +Everything that requires a licensed human
This is the billable work, and it stays yours.
We don't touch the minutes you bill. CPT 99457 requires interactive communication with the patient by your clinical staff. That stays with your team, it stays billable, and CMS's proposed 2027 requirement that the time come from a direct employee doesn't disturb it. What changes is what your employed clinicians spend those minutes on: the patients who need judgment, not the ones who needed a reminder.
We won't print a staffing multiplier, because we haven't measured yours and neither has anyone else selling you this. Once we're working together, we instrument patients-per-clinician at the billable threshold on your own book, before and after, and you keep the number whatever it says.
Out of range, checked, and routed to you.
An agent checks in.
A reading lands outside the range your clinician set. An agent reaches the patient and collects the first round of symptoms, in your program's voice, so nothing sits in a queue waiting for someone to have time.
Checked before it sends.
Every message the agent would send is checked against your program's rules first, then allowed, rewritten to stay in bounds, or stopped and routed to a person. The decision is recorded next to what was originally proposed, and models are reached only under BAA.
Surfaced, not decided.
What the patient said, the reading behind it, and what was proposed against what actually went out, all delivered into the record your team already works in. The agent surfaces; your clinician decides.
AI your clinicians trust. Governance your compliance team approves.
Routed only to models under BAA
Patient data only reaches models AnyBio holds a BAA with, never a non-compliant provider, and the routing is policy-driven and logged.
A policy on every output
Every message is checked against your program's defined wellness-or-clinical lane and allowed, rewritten to stay within bounds, or stopped for review.
Consequential calls surface to a human
Clinically consequential decisions surface to your clinical team rather than being made autonomously. The agent surfaces; it does not decide.
The record your customers will ask for
Every output is bounded in what it can do and fully logged, so the security review that gates adoption has one policy to examine, not one per tool. SOC 2 Type II.
You operate the program.
Keep the devices you've deployed. AnyBio adds governed agents over the monitoring you already run: they take the non-clinical outreach so your employed clinicians cover a bigger panel at the top of their license, and the result lands in the record system you already run. No rip-and-replace, and because the rails are neutral, it works across every device your program spans.
You run RPM for provider clients.
White-label the same governed agents under your own brand. One webhook on the measurements you already collect, then symptom outreach, adherence, and check-ins run program by program, each under its own rules and its own audit trail. Add it to your platform as a SKU your clients can turn on. You keep the workflow and the customer relationship; the layer underneath governs what software says to a patient and keeps the record of it.
The result lands where your team already works.
Native Epic adapter in development; EHR-ready FHIR and HL7v2 delivery are the live paths into record systems today.
What RPM operators ask us first
A working session with our team.
Bring the people who would run the program. We'll walk through your monitoring footprint, show what the agents would and would not be permitted to do, and scope what a first program looks like. The one question worth arriving with: when your AI writes something to a patient, what happens between the model and the send?
Put the work that doesn't bill on the rails. See what it frees.
We'll walk through your program, what the agents would and would not be permitted to do, and what your governance review actually has to cover. Your clinicians keep the calls that need a license; the rest runs on rails that already run in production.
