Ask the workspaceGrounded in your documents
Ask anything across your documents…
For physician owner-operators
Private AI desktop for clinics, medspas, and physician-owned practices.
A dedicated AI environment that helps your team search charts, draft prior auths, summarize notes, triage messages, and run the back office — without exposing PHI to outside AI providers.
Book a private AI systems call
For independent clinics, specialty practices, medical spas, and physician-owned groups handling protected health information and confidential operational documents.
The problem
You want AI in the practice. You cannot put protected health information into ChatGPT.
Clinics run on documents — charts, payer rules, intake forms, lab results, marketing templates. The work is repetitive and judgment-heavy. Public AI tools cannot touch most of it, so the work stays manual.
Trapped in
- Patient charts and visit notes
- Prior auth criteria and payer rules
- Lab and imaging results
- Portal messages and call logs
- Intake forms and consents
- Marketing templates and follow-up sequences
- Credentialing and licensure tracking
Eaten by manual work
- Drafting prior auths line by line
- Re-reading charts before a visit
- Typing the same patient replies again
- Pulling forward last visit’s plan
- Chasing missing intake information
- Following up on outstanding labs
- Sending the same medspa reminders
Today: scattered, manual, repeatedOne private layer over the same documents
A private AI desktop lets your team do this work faster, without a single chart leaving your network.
What you build
A private AI layer over your charts and your back office.
Designed to sit alongside the systems you already use. Pulls from the EHR, the portal, your document folders, and your operational templates. Everything happens inside your perimeter.
01
Search across charts and visit notes
Plain-English questions over your patient documents.
02
Draft prior authorizations from the chart
Step therapy, dx, labs, and payer criteria pulled together.
03
Summarize labs, notices, and correspondence
A structured brief instead of a stack of PDFs.
04
Triage portal messages
Drafted replies in your clinical voice, ready to send.
05
Pre-visit chart prep
Today’s schedule with the relevant chart context inline.
06
Operational drafting
Letters, refusal templates, referral notes, internal memos.
07
Practice and medspa marketing
Treatment follow-ups, recall lists, and intake nudges.
08
Credentialing and renewal tracking
DEA, state license, CME, malpractice, payer credentialing.
PHI stays on the rig. Your staff get serious AI help on the work that previously had to stay manual.
Principles
How we work, in four lines.
These constraints are baked into the architecture of every clinic and medspa deployment. Not policies we hope to honor — design decisions.
The AI runs in your building.
Models execute on the on-site rig. Patient data is not transmitted to a cloud inference endpoint. No vendor BAA on the inference path.
Human approval on every write.
Every prior auth, patient reply, and clinical draft requires explicit physician sign-off. The agent prepares, you decide. Nothing leaves under your name without you.
Voice trained on your real letters.
Calibrated on 15–20 of your real PAs, referral letters, and portal replies before anything goes out. The assistant writes in your cadence and clinical tone.
Air-gappable by design.
The rig can run fully offline. Cloud is opt-in for specific non-PHI workflows. Default posture is local, encrypted at rest, and yours.
Example deployment
Independent practice AI deployment
A reference design built for physician-led practices. Sized for a partner-run clinic rather than a hospital system. The same architecture scales cleanly to medspas, multi-provider groups, and specialty practices.
Project sizeTens of thousands
ClientPhysician-owned primary care practice
Use caseClinical assistance + back-office automation
A dedicated AI rig sits in the practice. Models, document processing, and chat all run locally. EHR, lab inbox, and patient portal connect through existing APIs. Nothing is written back into a patient chart without explicit physician sign-off.
Document types included
- Patient charts and visit notes
- Lab and imaging results
- Prior auth payer criteria
- Portal messages
- Intake forms and consents
- Treatment plans and marketing templates
- Credentialing and licensure documents
- Internal SOPs and protocols
Core workflows
- Morning brief — schedule, labs, refills
- Prior auth drafting
- Portal message triage and reply drafts
- Pre-visit chart prep
- Refill review
- Referral letter drafting
- Medspa follow-up and recall
- Credentialing and renewal alerts
Security model
- On-device inference — no cloud BAA required
- Patient data stays on the rig
- Human sign-off on every write
- Role-based access by clinical role
- Audit logging on every prompt
- Voice capture off by default in exam rooms
- Air-gappable for the most sensitive workflows
Have a similar workflow?
Book a call →
Reference deployment · physician-owned primary care practice
Who this is for
Built for physician owner-operators.
This is for the partner who is personally on the hook for both the medicine and the operations. Where the patient relationship and the practice’s reputation are not separable from the work.
Strong fits
- Independent primary care practices
- Specialty practices (cardiology, derm, ortho, GI)
- Medical spas and aesthetic practices
- Concierge and direct-pay clinics
- Multi-provider physician groups
- Surgery centers
- Behavioral health practices
- Owner-operated specialty clinics
Common patterns
- Owner-operator wears multiple hats
- Staff are the bottleneck, not the model
- PHI rules out public AI tools
- Prior auths and inbox messages dominate the day
- Care continuity depends on prior visits
- Marketing and recall sit on someone’s desk
- Credentialing deadlines slip
- Senior judgment trapped in old charts
Especially good fit if you hear
- “My MA is drowning in portal messages.”
- “We spend two hours a day on prior auths.”
- “I can’t put a patient note into ChatGPT.”
- “I review every chart the morning of — there’s no time.”
- “Our medspa follow-up is whoever has a minute.”
- “I keep meaning to fix this but I’m in clinic.”
In their words
What physicians and practice managers tell us.
“My MA went from drowning in portal messages to clearing the inbox before lunch. The replies are in my voice — I sign off, but I am not the bottleneck anymore.”
“Prior auths used to eat half a day. Now the letter is drafted from the chart by the time I think to ask for it. The compliance posture is the version my counsel can actually sign off on.”
“We run a busy medspa. The follow-up cadence used to be whoever had a minute. Now the right client gets the right message on the right week, in our voice.”
Deployment
A rig in your practice. Models running locally.
We deploy a dedicated AI workstation to the practice. Models, context, and document processing all run on-site. Cloud is opt-in for specific non-PHI workflows. Default posture is local, encrypted at rest, and yours.
Inputs
Patient charts and visit notesPrior auth criteria and payer rulesLab and imaging resultsPortal messages and call logsIntake forms and consents
Your perimeterNothing leaves your network
Private AI desktopLocal models · grounded retrieval · approval gate
Document indexLocal modelsPermissionsAudit log
Outputs
Search across charts and visit notesDraft prior authorizations from the chartSummarize labs, notices, and correspondenceTriage portal messagesPre-visit chart prep
The rig
A quiet AI workstation sized to your practice — a Mac Studio or a small GPU box.
On-device inference
Local Llama, Qwen, and Whisper models. Patient data is not transmitted to a cloud inference endpoint.
EHR and portal integration
Reads through your existing APIs. We work with your EHR vendor for connection setup.
Voice capture
Optional. Mobile, tablet, laptop, wearable, and exam-room mics. Off by default in rooms.
Human approval on writes
Every PA, patient reply, and clinical draft requires explicit sign-off before it goes out.
Maintained software
We update the layer through a secure channel. No vendor access to patient data.
A one-time capital cost for the rig. No per-token charges. No cloud dependency. No vendor that could change their terms next quarter.
What happens on the call
Book a 30-minute private AI systems call.
We talk through your actual practice — what the staff do, where the day goes, and where AI would actually help. No clinical sales theater.
We will map
- Where your day and your staff’s day actually goes
- Which workflows are PHI and which are not
- What the EHR can and cannot reasonably feed
- What we would deploy in Phase 1 and Phase 2
- What the rig and the build would cost
- Whether a private AI desktop makes sense for your practice
Book a private AI systems call
You leave with a rough architecture, a priority workflow map, and an honest view of whether this is worth building.
Project packages
Three ways to engage.
Most practices start with an audit or go straight to a Phase 1 build. The right starting point depends on how clear the workflows are.
Workflow audit
For practices unsure where AI helps
We spend two to four weeks watching how the practice actually runs. You leave with a buildable architecture, a priority list, and an honest go/no-go.
Phase 1 build
For practices ready to deploy
A working AI desktop — morning brief, prior auth drafting, credentialing tracker, and the foundational chart and portal integrations. Live in the first 30 to 60 days.
Full build + automation
For multi-phase deployments
Phase 1 plus inbox triage, voice capture across mobile and exam rooms, back-office automation, and a longer-horizon plan for billing watch and practice growth.
Phase 1 commonly lands in the tens of thousands depending on practice size, EHR, document volume, and security requirements. We share a fixed-scope SOW before any commitment.
Security
Built for clinical work.
PHI lives on the rig. The AI runs on hardware in your building, on your network. No vendor BAA on the inference path. No third-party processing of patient data.
01On-device inference — no cloud BAA required
02Patient data is not transmitted to a cloud inference endpoint
03Human sign-off on every PA, reply, and clinical draft
04Role-based access by clinical role
05Document-level permissions on charts and notes
06Audit logs on every prompt and retrieval
07Voice capture off by default in exam rooms
08Air-gappable for the most sensitive workflows
09Local Whisper transcription — audio is not uploaded
10Admin kill-switch and retention controls
The default posture is local, encrypted at rest, and yours. Cloud is opt-in for specific non-PHI workflows.
Who builds this
Founder-led. No clinical sales theater.
We work hands-on, founder-led, end-to-end. The same depth we have brought to Fortune 100 environments, scaled to fit a single physician-led practice. The people who design the system are the people who ship it.
BJ
Brian Joseff
Co-founder
Designs the on-device architecture and the security model. Background in private-cloud AI for regulated industries.
CS
Chandler Supple
Co-founder
Leads on-site delivery with each practice. Background in operational AI for high-volume document and inbox work.
FAQ
Common questions from practice owners.
The deployment is designed to keep PHI inside your perimeter. Models run locally, audio is transcribed on the rig, and patient data is not transmitted to a cloud inference endpoint. We work alongside your compliance officer or counsel on policies, BAAs for the systems you keep cloud-based, and an audit trail you can produce on request.
Build a private AI desktop for your practice.
For physician owner-operators with charts to read, prior auths to draft, and a back office that does not run itself.