Practices and firms

Clinics: help at the front desk, nothing clinical automated

For practice managers and front-desk leads at dental and allied-health clinics with one to three locations. Workflows prepare replies, answers and requests; your staff decide.

Your week, in your words

iSystematic works with clinic front desks whose week goes on missed calls and messages, the same ten questions, confirmations, recalls, no-shows, insurance questions and review requests.

Sara (fictional), a front-desk lead at a dental clinic, starts each morning with messages from the phone line, the web form and email, and answers most of them from memory. The workflows on this page are written for a week like hers.

Where most start

Most clinics start with a setup session for two workflows: W02 Inbox Triage and Draft Replies and the W03 Approved FAQ Assistant, configured with you in your own AI assistant account. Messages are sorted and drafted; your staff send.

A different start fits clinics rebuilding their website and patient information together: the Practice Platform, with nothing clinical automated.

What we set up

Each workflow prepares work for a person to check. Status on 9 October 2026: free cards in testing, packs proposed. Workflows without a link are designed and not yet published.

W01·Free card: In testing

Daily Briefing

A dated morning briefing of today's commitments, priorities and open questions, from your own notes.

Read more →
W02·Free card: In testing

Inbox Triage and Draft Replies

Sorts messages into scheduling, billing, general, complaint or urgent, and drafts replies from approved facts only.

Read more →
W03·Free card: In testing

Approved FAQ Assistant

Answers only from your approved FAQ, cites the section, and says when staff must confirm.

Read more →
W04·Not yet published

Appointment Request Preparation

Prepares booking options and a confirmation draft. Nothing changes until staff approve.

W16·Free card: In testing

Invoice Reminder Preparation

Prepares reminders for confirmed unpaid invoices only, with the exact amounts supplied.

Read more →
W17·Not yet published

Feedback Themes and Review Requests

Summarises feedback, sends complaints to staff, and drafts one neutral review request, never filtered by rating.

W19·Not yet published

Weekly Owner Review

A weekly memo from figures you supply, with missing data shown and at most three recommended actions.

W21·Not yet published

Public Policy Watch

Checks official sources you name and reports real changes with links. A failed source is reported, never read as no change.

Works with your systems

Most of this runs without touching your practice management system; confirmations and recalls are the exception, and we say so.

What it doesWhat it needs
Sorting messages, drafting replies from your FAQ, preparing appointment requestsNo integration. It runs in your own assistant account and does not touch your practice management system.
Confirmations and recallsSchedule data. That is a custom integration (level 5), assessed with your own practice management system first.

Software is named here only to say what has been tested with it. Nothing on this page implies an integration that has not been built.

How it is built

Every system we build for a clinic keeps the promises below; each is the plain-words form of a framework explained on How we build.

What we promiseWhat stands behind it
It never promises a patient anything you have not approvedThe Boundary Invariant
It checks every answer against your own sources before anyone sees itPEVG: the verifier
It cannot change its own instructions; you approve what it learnsPARA: adaptation bounded like action
Nothing goes live until a named person at your clinic signs it offFive-Gate Deployment Model™: G3 Approval
Confidential questions go only to the AI services you approved, in Canada where requiredGoverned Routing Policy; Cross-Border AI Architecture Patterns™
It has a monthly spending limit it cannot exceedBudget as Boundary
We check every service we put into your setupAI Vendor Risk Framework (AVRF)™
If something goes wrong, we can show exactly what happened, and fix the causeAI Incident Response Protocol (AIRP)™ and the event identity
You can see what it did and whyThe BOE Declaration

These promises describe the systems we build and run for you. A free card is a manual routine: you run it and check its output yourself.

What always stays with you

Your staff decide these, every time.

  • Any clinical question
  • Any booking change
  • Anything sent to a patient
  • Any refund or fee decision

What we never do

We never build any of these into a clinic's setup.

  • Clinical advice or triage
  • Patient records, charts, X-rays or health numbers in consumer AI accounts
  • Review gating: asking only satisfied patients for reviews
  • Writing or posting reviews on anyone's behalf
  • Booking changes without staff approval

What the research says

This is one published study, not our result. It measured physicians writing clinical notes with an AI scribe, not clinic front desks.

About 9.5% less
Physicians spent about 9.5% less time writing each note with one of two ambient AI scribe tools (Nabla); the other tool (DAX) showed no significant change.
238 outpatient physicians across 14 specialties, randomised 1:1:1 to DAX, Nabla or no scribe; 72,369 visits over two months at one academic institution, measured as time-in-note per note in the Epic health record.
Single centre; physicians, not front desks; English-only visits. Burnout and workload were secondary measures and were not statistically tested; the authors say any improvement needs confirmation in larger trials. Clinicians reported that notes occasionally contained clinically significant inaccuracies. The scribes were used in only 29.5% (Nabla) and 33.5% (DAX) of visits, and time-in-note leaves out editing inside the vendors' apps.
Lukac et al., NEJM AI, 2025 ↗

Published studies, not our results. Ours will come from documented pilots.

Sara is a fictional persona. Workflows prepare; people decide.

Start

Tell us the task that fills your front desk's week

We reply with one recommended start.