Request a call

AI phone, intake and recall systems for clinics

Front desks in busy practices spend the day on the phone and the evening on paperwork. We build voice agents and intake automation that take the routine part of both, and hand anything clinical to your staff.

Request a 30-minute scoping callSee the use cases

Expect a reply within 1 working day, with times for the call.

Five concentric glass rings spreading out like ripples, thinner and fainter toward the outside, with a tiny clear bead at their center and a glowing orange bead resting on the inner ring.

In short

For medical, dental and specialty clinics, AI pays off first at the front desk: a voice agent that answers after-hours calls and books appointments, intake forms read into the practice management system and referral letters sorted for coordinators. In the US, The Method Crew only puts a vendor in the patient data path once it has signed a business associate agreement, and no model gives clinical advice.

Updated

Where AI and automation pay off first

Each use case names the systems it connects to and the services that build it. Every service page lists its timeline and price band.

  • After-hours call answering and booking

    A voice agent answers on your existing number through Twilio, books, moves or cancels appointments against the live schedule and texts a confirmation. Any caller describing an emergency hears your emergency instructions word for word, as your clinicians wrote them, and is offered your on-call line.

    Service

  • Intake form extraction

    New patient forms, insurance cards and history questionnaires are read and the fields are prepared for staff to confirm in Dentrix, Open Dental, athenahealth or whichever system you run. Nothing is saved to the patient record until a person approves it.

    Service

  • Referral letter sorting

    Referrals arriving by fax, email or portal are matched to the patient, sorted by referring provider and requested service, and placed in a coordinator's queue with the key details pulled out. Any letter the referrer marked urgent goes straight to the top.

    Services

  • Recall and reminder sequences

    Hygiene recalls, annual checkups and unconfirmed appointments trigger text and email reminders from your practice management data. This is rules-based automation. No model is involved.

    Services

  • Front-desk knowledge assistant

    An internal assistant over your procedures, payer notes and fee schedules, so a new receptionist can find the answer with a link to the source document instead of interrupting the practice manager.

    Service

  • Review requests after visits

    After a completed visit, the patient gets one review request by text. Replies are summarized weekly for the practice manager, and no public response ever mentions a patient's health or treatment.

    Service

What shapes the build

The rules, risks and sign-offs we settle with you in the first 2 weeks, before any real data moves.

  • Business associate agreements before patient data

    In the US, a vendor that handles protected health information on your behalf generally needs a signed business associate agreement (BAA). We list every vendor in the data path, confirm each BAA with you before the build and keep patient data out of any tool that will not sign one. Your privacy officer makes the final call.

  • We do not call any system HIPAA compliant

    Compliance belongs to your practice and depends on your policies, training and every system you run. We document the controls we build (access limits, audit logs, retention settings) so your privacy officer or compliance advisor can assess them against HHS guidance.

  • No clinical advice from a model

    The voice agent books, reschedules and answers practical questions about hours, parking, forms and insurance accepted. It does not assess symptoms, interpret results or give medical advice, and its scripts are approved by a clinician before launch.

  • Clinics outside the US

    Health data is special category data under the GDPR and UK GDPR, and other countries set their own rules. Point us to your regulator's guidance and your data protection officer, and we build to the requirements they set.

A clinic waiting room with a long sofa under a framed frosted-glass panel, and a slatted front desk with a wall phone on the right.
The waiting room and the front desk, where calls, intake forms and recalls start. Anything clinical stays with your staff.

Where should a clinic start with AI?

Start with the phone. Pull 2 weeks of call reports from your phone system and count the calls that rang out, went to voicemail or came in after hours. For many practices that number is the clearest case for a voice AI agent, because each missed call is a booking that may go to another clinic.

The first version should do one job well: answer, book or reschedule against your real schedule, and take a message for anything else. The AI receptionist blueprint shows the call flow, the handoff rules and what the front desk sees the next morning.

Whether it can book directly depends on your practice management system. Some expose scheduling through an API, some through an integration partner, and some only through exports. Where direct booking is not possible, the agent captures the request and texts the patient a booking link, which is less elegant and still far better than voicemail.

Intake comes second. Document processing on new patient forms and referral letters saves typing every day, but it touches the patient record, so it needs a careful review step.

Recall reminders need no AI. If your recall list is the gap, we build plain business process automation on your existing data, which costs less to run and is easier to audit. Our comparison of AI agents, workflow automation and chatbots explains where that line sits.

How do patient data protections shape the build?

Every design decision starts from one question: which vendor sees which piece of patient information, and has it signed a BAA with you.

In practice that means:

  • a written data map listing every vendor in the call, intake or referral path, with the BAA status of each, agreed before we build
  • the minimum patient data each step needs, so a booking step sees a name and appointment slot, not a history
  • audit logs of every action the system takes, stored in an account you own
  • staff approval before anything is written to the patient record, a human-in-the-loop step that stays in place after launch
  • retention settings agreed with your privacy officer, including how long call recordings and transcripts are kept

HHS publishes the rules on covered entities and business associates and the required elements of a business associate contract. Read them with your compliance advisor. We do not give legal advice, and we do not label a system compliant. We hand you the documentation to judge it.

Two more limits apply to every build. The agent tells callers it is automated and offers a person. And patient-facing scripts are written by your staff or approved by a clinician before launch.

Staff use of public AI tools needs rules as well, because a pasted referral letter is patient data. Our AI acceptable use policy template treats health information as restricted data that only approved systems may receive.

What does a sample 90-day plan look like?

This plan suits a practice with 1 to 5 locations starting with an after-hours voice agent.

WeeksStageWhat happens
1 to 2MapCall reports, front-desk interviews, a check of your practice management system's scheduling access, the vendor and BAA list, and the acceptance test for the agent
3 to 4ProveThe agent takes test calls from your staff using real call scenarios, books into a test schedule, and is scored against the agreed bar
5 to 7BuildLive scheduling, emergency handling, consent wording, logging and the morning summary for the front desk, with a demo every week. After-hours calls go live in week 6 or 7, with staff reviewing every call for the first 2 weeks
8 to 13RunDaytime overflow follows if you want it, intake automation can start as a second project, and the first monthly report covers calls, bookings, handoffs and cost

If the agent cannot handle the test calls reliably by week 4, we stop and you keep the findings.

This is the full receptionist from the AI receptionist blueprint: booking, answers from approved scripts, messages and transfers. The build costs $12,000 to $30,000 depending on locations, languages and integration depth. Per-minute telephony and voice usage is passed through at cost. A care plan from $1,500 a month covers monitoring, script updates and fixes when a vendor changes something.

Services we would start with

Timelines and starting prices for the services that do most of this work.

See all 20 services

Blueprints that fit

Reference designs with the flow, the stack, the guardrails, the build time and the price.

See all 8 blueprints

Questions about AI in healthcare and clinics

The full FAQ covers pricing, ownership, data and timelines.

Is your AI receptionist HIPAA compliant?

We do not label any system HIPAA compliant, because compliance depends on your practice's policies and every vendor involved. What we do: use only vendors that sign a BAA with you for patient data, limit what each step sees, log every action and give your privacy officer a written list of vendors and controls to review.

Can the voice agent book directly into our schedule?

If your practice management system allows scheduling through an API or an integration partner, yes. We check this in the first week for your exact system and version. Where direct booking is not available, the agent captures the request and texts the patient a booking link, and your front desk confirms it in the morning.

What happens when a patient calls with an emergency?

The agent does not assess symptoms. If a caller describes anything your clinicians list as urgent, or simply asks for help, the agent reads your emergency instructions word for word and offers your on-call line. Those scripts are written or approved by a clinician before launch and tested on every release.

Will patients know they are talking to an AI?

Yes. The agent says so in its greeting, and callers can ask for a person at any point. During opening hours that transfers to the front desk. After hours it takes a message for a callback. Disclosure rules for automated calls vary by state and country, so we set the wording with your advisor.

How long does it take to go live?

In our standard plan, after-hours calls go live around week 7, after 2 weeks of mapping, 2 weeks of testing on staff calls and a build of about 3 weeks. Practice management systems without scheduling access add time. Intake and referral automation can follow as a second project of 6 to 12 weeks, the usual span for a document processing build.

When is AI the wrong choice for a clinic?

Recall reminders, appointment confirmations and review requests are rules, so we build them without a model. AI is also wrong for anything that needs clinical judgment, such as triaging symptoms or explaining results. And if a vendor in the path will not sign a BAA, patient data does not go to it, whatever the feature list says.

Get a fixed price for your first build

Tell us on a 30-minute call where the hours go. We'll say whether it needs AI, plain automation or neither.

Request a 30-minute scoping call

Within 5 working days of the call you get a written scope and a fixed price.