The AI workforce for medical practices
Thirteen specialised agents are being built to handle calls, scheduling, documentation, prior auth, denials, posting and credentialing inside the EMR you already use. Every action is logged, reviewable and reversible — and when confidence drops, a certified human picks it up.
13
Specialised agents
Purpose-built per workflow, not one chatbot doing everything
100%
Actions logged
Who, what, when, why and the source record for every step
< 2 hrs
Human review SLA
Anything below threshold lands in a staffed queue
0
PHI used for training
Contractual, in writing, with every model provider
How it runs
A call, text, fax, referral, portal message or claim response lands in the practice.
Routing is by channel and intent, not by keyword. The agent reads context before it acts.
Books in the EMR, files the document, submits the claim, posts the payment — through the integration layer.
Inputs, decision, action, confidence and version are written to the evidence layer.
Below-threshold work goes to a review queue with a two-hour SLA during business hours.
The agent catalogue
Each agent has one job, its own guardrails and its own accuracy reporting. Deploy one, or deploy the workforce.
Every inbound channel answered, routed and closed — with a transcript a human can pick up.
Answers every call on your existing number, day or night. Books and moves appointments in your EMR, captures refill requests, answers routine questions from your own practice information, and leaves staff a transcript plus a task.
Turns website visitors into booked appointments. Answers questions from your own practice information, checks provider availability, and hands the visitor a real appointment instead of a contact form.
Two-way text that confirms, reminds and refills the schedule. Handles replies like 'can I move it to Thursday', keeps a TCPA-compliant opt-out record, and fills cancellations from your recall list.
Triages the practice inbox: sorts patient questions, payer notices, vendor noise and referrals, drafts replies for staff approval, and files what belongs in the chart.
Reads inbound faxes, scans, portal messages and referrals, identifies the document type, matches the patient, files it in the right chart and turns orders into tasks.
From conversation to a signed, coded, filed note without the provider typing.
Listens to the visit and drafts the note in your format — SOAP, APSO or your own template — with suggested codes, orders and follow-up. The provider reviews, edits and signs.
Turns the signed note into codes and charges, checks them against payer rules and your fee schedule, and queues anything ambiguous for a certified human coder.
Intake, eligibility and recall handled before the patient is in the room.
Collects history, consents and insurance before the visit, verifies eligibility and benefits in real time, and writes clean structured data into the right EMR fields.
Works your recall list and quality gaps continuously: overdue screenings, missed follow-ups, annual visits and chronic care check-ins — outbound, in your voice, with the appointment booked.
Prior auth, denials, appeals and posting worked continuously instead of in batches.
Checks whether a service needs authorization, assembles the clinical packet, files it, and tracks every request to a decision — with the deadlines that matter.
Works deny-rail and clearinghouse rejections continuously: categorises each denial, drafts the appeal with clinical support attached, and escalates the patterns that keep recurring.
Reads ERAs and EOBs, matches payments to claims, posts insurance and patient amounts, verifies contractual adjustments, and reconciles every posting against the bank deposit.
Credentialing and expirables tracked to the day, with an audit trail.
Keeps every provider's credentials, CAQH profile, payer enrolment and expirables current — prefilling applications, chasing payers and escalating before anything lapses.
Why this is different
Plenty of products will answer a phone. Ours hand the hard parts to certified people and prove what they did.
Escalations do not land in a ticket queue. They land with the certified coders, billers, credentialing specialists and auditors we already staff for your practice. The agent is the front door; the humans are the safety net.
Choose per action whether the agent commits or queues for approval. Start with everything queued, watch the accuracy numbers, then widen autonomy workflow by workflow. Nothing goes straight to production unreviewed unless you say so.
Every decision carries its inputs: the transcript segment, the chart field, the payer rule, the fee schedule line. Export the full audit trail as CSV or JSON whenever you want, including after you leave us.
FHIR and HL7 where the vendor supports it, secure browser automation and structured document intake where they do not. You keep your system, your tax ID and your contracts. We do not sell or resell software.
We sample agent output every month and score it against the same standards our audit team applies to human coders and billers. Accuracy by workflow, by agent, by provider — reported, with the misses shown.
Agent configuration, prompts, routing rules and logs are yours and exportable in a standard format. Pricing is per agent or per provider. If you leave, you leave with the configuration.
Three real workflows
Patient access
Documentation
Revenue cycle
Platform
Phone, SMS, web chat, email, fax, portal and in-room capture — every way a patient or payer reaches you, normalised into one work queue.
Thirteen specialised agents, each with its own scope, tools, confidence thresholds and escalation rules. No single agent holds the whole practice.
FHIR, HL7 v2, secure RPA and document intelligence connect agents to your EMR, PM, clearinghouse, lab and payer portals.
A human work queue with SLAs, staffed by our certified team or yours. Approvals, corrections and overrides all feed back as training signal.
Immutable audit log, reasoning trace, model and prompt version, reviewer identity and outcome — exportable on demand.
Governance
Autonomy is granted per workflow, measured, and revocable. These six rules are not configurable — they are architecture.
The 30-day pilot
One workflow, one measured baseline, one written report. We deploy a single agent with everything queued for human approval, measure it against your own numbers, and tell you whether the maths works — including when it does not.
Pilot slots are limited because each one is configured by hand. If we cannot see a credible return in your volume, we will say so before you pay for anything.
FAQ
Pick the workflow that hurts most — the phone at lunch, the fax tray, the denial queue. We will show you the agent on your own scenarios, and tell you what we would not automate.
A 30-minute call with an operations lead. No obligation, no sales sequence, and we will say so if you do not need us.