How to Automate CharmHealth

CharmHealth carries the whole outpatient day

CharmHealth is where an outpatient day gets written down. It puts the electronic health record, practice management and medical billing on one platform, and it is used by independent clinics, larger groups and health systems. Plenty of people still search for it under its older name, ChARM EHR.

An outpatient practice is a queue with a memory. Patients arrive at fifteen-minute intervals, and each one leaves the room having set something in motion: a prescription sent, a scan ordered, a referral written, a course of treatment to be reviewed in six weeks. The visit is over in minutes. What the visit started can take a month to resolve.

CharmHealth records all of it faithfully. The order is in the chart, the appointment is on the calendar, the claim goes out on time.

What it cannot tell anyone, on its own, is what happened next.

Every order leaves something open behind it

A practice's real workload is not the visits. It is the tail behind them.

The prior authorisation submitted on Tuesday that the payer has not decided. The CT ordered three weeks ago whose report never came back from the imaging centre. The referral to a specialist, with nothing in the chart to say whether the patient ever went. The immunotherapy vials that run out this month. The patient told to return in six weeks who has not rebooked. The claim denied in January whose appeal window closes in March.

None of that is in CharmHealth, because none of it happened inside CharmHealth. It sits in an imaging centre's portal, a payer's authorisation site, a specialty pharmacy's status page, each behind its own login.

So somebody at the practice keeps a list. Sometimes on paper, sometimes in a spreadsheet, frequently only in their head, worked through in whatever gaps the day leaves. When that person is out, the list is not worked.

CharmHealth

Prior authorisations followed daily Each pending request checked on the payer's own site, so an approval reaches scheduling the day it lands.

Results matched to the order Reports pulled from the imaging centre's portal and set against the order that asked for them in the first place.

Recalls that do not rely on memory Who is due back, who never rebooked and whose refill runs out this month, prepared as a list a person can work.

CharmAI Scribe writes the note, not the follow-up

CharmHealth automates a great deal of the visit, and it is worth being precise about how much.

CharmAI Scribe drafts the clinical note from the encounter, and CharmCopilot acts as an assistant across documentation and workflow. SOAP Templates mean a routine visit is largely written before the physician starts typing. Realtime Eligibility Checks verify coverage in real time, Electronic Claims go out, and Electronic Remittance Advice comes back into the same system that raised the invoice. Referral Management, e-Prescribing and Lab Integration sit inside the chart rather than in a separate program.

Every one of those works on something CharmHealth already holds, or on a channel it was built to receive. Eligibility comes back because a route exists for it. The remittance arrives because a route exists for it.

The routes that do not exist are the ones the open list depends on. Referral Management sends the referral; it cannot say whether the specialist saw the patient. The chart holds the CT order; the report is on the imaging centre's website. The prior authorisation went into a portal that will not be announcing when its status changes. Nothing is wrong with the software. The answer is simply on the far side of somebody else's login.

The loop gets closed from outside the chart

The open list is the practice's real workload, and it is held together by whoever remembers to check it. That can be a schedule instead of a memory.

Every prior authorisation the practice is waiting on, checked against the payer's own site each morning, so an approval reaches scheduling the day it lands and a stalled one is chased on day four rather than on the morning of the procedure. The imaging centre's portal read daily, with new reports pulled down and set against the order that requested them, so nobody discovers a three-week-old result at the next visit.

Referrals followed up on the receiving practice's system, so an appointment that was never made is visible while it still matters. Denials collected from each payer portal with the appeal deadline attached, sorted by which one closes first. Recalls and refills turned into a list somebody can actually work: who is due, who has not rebooked, whose immunotherapy runs out this month.

One line here does not shift, and everything else depends on it holding. Gathering runs unattended, all night if you like. Anything that changes a clinical or billing record, and anything that reaches a patient, is prepared and waits for a qualified person to read it and send it. The point is to remove the fetching, not the judgement.

The waiting can be watched by something else

Every morning, the same portals get opened to see whether anything changed since yesterday. There is no clinical reason a person has to be the one doing it.

WebRun is an AI agent that drives a real Chrome browser, signed in the way your staff sign in. It opens the payer site, the imaging centre, the pharmacy portal or a referring practice's system, reads what is there, and brings it back as a worked list with dates attached.

It runs on your schedule inside your own private environment. Sessions are not shared between workflows, a workflow can be locked to a named list of domains, and you can watch a run and stop it mid-way.

The workflows below are already built, and each one names exactly what it opens.