AI agents that respect your patients' data.
Chatbots, intake automation, and scheduling AI — built and audited the same way we build everything else: under your BAA, with PHI handling reviewed at every step. Separate engagement from your monthly compliance plan.
What We Build
We build AI agents that handle the repetitive work around a healthcare website without putting PHI in places it shouldn't be.
Patient FAQ Chatbot
Answers common questions about services, hours, insurance, parking, and prep instructions — 24/7, in your tone, with safe handoff to staff for anything clinical.
Smart Intake & Triage
Collects structured intake before the visit, flags red flags for staff review, and writes a clean summary into your practice management system instead of a wall of free text.
Scheduling Assistant
Books, reschedules, and confirms appointments through your existing calendar. Handles eligibility checks, reminder workflows, and waitlist fills without staff time.
Internal Knowledge Bot
Lets your team query your SOPs, billing rules, compliance policies, and provider handbook in plain language — fewer interruptions, faster onboarding for new staff.
How It Stays HIPAA-Compliant
Hosted on HIPAA-eligible models
We use AI providers that sign a BAA — Anthropic, AWS Bedrock, Azure OpenAI Service, or self-hosted models on HIPAA-compliant infrastructure. Never the consumer-grade API tier.
PHI handling audited at every step
Before launch we trace every place the bot can touch patient data, verify it's encrypted in transit and at rest, and document the data flow for your compliance file.
Conversation audit trail
Every conversation is logged with timestamps, retention controls, and access restrictions — so a HIPAA audit or breach review has a clear record to work from.
Safe handoff to staff
The bot is built to know its limits. Anything clinical, any red flag in intake, any request outside its scope — it routes to a real person with full context, instead of guessing.
Pricing
AI implementation is a separate engagement from your monthly compliance plan — project work plus ongoing maintenance, not a subscription tier.
Starter Bot
FAQ assistant or single-purpose bot. Right for solo practices testing AI before committing.
- 1 use case (FAQ, intake, or scheduling)
- Trained on your site content & uploaded docs
- Hosted on a HIPAA-eligible model with BAA
- Audit log of every conversation
- Staff handoff for anything clinical
- Quarterly accuracy review
Practice Assistant
Multi-purpose assistant integrated with your practice management software. Built for Group Practice tier and above.
- Up to 3 use cases combined
- Practice management integration (Dentrix, Jane App, ChiroTouch, WebPT, etc.)
- Structured intake → EHR workflow
- PHI-aware response filtering
- Monthly conversation review & retraining
- Custom escalation rules per provider
Custom AI Workflow
Multi-step automations across EHR, billing, and patient comms. For multi-site networks, DSOs, and health systems.
- Custom architecture & multi-system orchestration
- Cross-location data flows with audit trails
- Dedicated model fine-tuning where appropriate
- Co-designed with your Privacy Officer & IT
- SLA-backed maintenance & retraining
- Quarterly executive review
Upfront covers scoping, build, compliance audit, and launch. Monthly maintenance covers conversation review, retraining, model updates, and minor adjustments. Third-party AI service costs (Anthropic, AWS Bedrock, etc.) are billed at cost. Heavy redesign or new use cases scoped separately.
How an Engagement Works
Discovery call
Free 30 minutes. We map the workflow you want to automate, the data it touches, and whether AI is the right tool — sometimes it isn't.
Scope & quote
Within a week: written scope, fixed upfront price, and a monthly maintenance estimate. No surprise change orders mid-build.
Build & compliance audit
Two to six weeks. We build, test against your real workflows, and run a PHI-handling audit before anything sees a real patient.
Launch & maintain
Monitored monthly. We review conversation logs, retrain on edge cases, and adjust escalation rules as your practice evolves.
Frequently Asked Questions
Why upfront plus monthly — why not a flat subscription?
Building an AI agent is real engineering work — scoping, training, integration, and a compliance audit. That's the upfront. Monthly covers the ongoing reality of AI: models change, edge cases surface, conversations need review. A flat subscription either underprices the build or overprices the maintenance.
Will the bot leak PHI?
We design every bot so PHI only flows through HIPAA-eligible providers under BAA, and conversation logs are encrypted and access-controlled. Before launch we run a PHI-handling audit — same scope as the technical compliance work in your monthly plan. Bots that don't pass don't launch.
Whose AI are you actually using?
Depends on the use case. Most builds run on Anthropic Claude (via the API tier that supports BAA), AWS Bedrock, or Azure OpenAI Service. For higher-sensitivity workflows we'll deploy a self-hosted model on HIPAA-compliant infrastructure. We pick during scoping based on your workflow, not a fixed vendor.
What about hallucinations?
We constrain the bot to your actual content (your site, your SOPs, your uploaded docs) and explicitly tell it to say "I don't know — let me get a human" instead of guessing. We also flag high-confidence-but-wrong answers in monthly review and retrain on them. This is why monthly maintenance matters.
Do we need a separate BAA for the AI?
Yes — every AI provider that touches PHI needs its own BAA, separate from your hosting BAA. We sign and manage these as part of the build, then track them alongside the BAAs in your monthly compliance plan.
Can we use our existing chatbot tool instead?
Sometimes — if the tool offers a BAA and you accept its limits on customization and audit. We'll tell you honestly during discovery. A lot of off-the-shelf chat widgets either don't support BAA or store conversations on infrastructure that wouldn't survive a HIPAA audit.
What if we want to cancel monthly maintenance?
30 days notice. The bot keeps running — you own the configuration, the prompts, and the integration code. What you lose is the conversation review, retraining, and model-update work, so accuracy will drift over time. You can re-engage later for a tune-up at hourly rate.
Start With a Free Discovery Call
30 minutes, no commitment. We'll tell you whether AI is the right fit for the workflow you have in mind — and if not, what is.
Schedule a Call