Method and sources

How the officers read your numbers.

Why it exists. Most practices are too small for a finance chief, an operations manager who reads the reports, or a marketing lead, so the owner does all three at night or not at all. The rules here come from published dental-industry surveys and practitioner summaries. They are written down below so you can judge them yourself. It is software only; there is no human coach.

The officers are software applying written rules to the numbers you give them. No AI model writes the reports, and none of your numbers is sent to a server. This page lists the ranges, where they come from, and how each action is chosen, so you can check any of it.

The AI Finance Officer

It sorts last month into a few groups (team, associates, lab, supplies, rent, marketing, everything else), divides each by collections, and compares the share with a typical range for a general practice in your region. Your own pay, loan payments, interest, depreciation and taxes are left out of overhead.

Line, % of collectionsUS overallCoastal California*Inland California*
Team24 to 30%28 to 35%26 to 32%
Lab5 to 10%5 to 10%5 to 10%
Dental supplies5 to 8%5 to 8%5 to 8%
Rent and occupancy4 to 8%6 to 10%5 to 8%
Marketing2 to 5%2 to 5%2 to 5%
Everything else8 to 14%9 to 15%9 to 15%
Total overhead55 to 65%60 to 70%58 to 67%

* California team and rent ranges are our own estimate from state wage and medical-office rent data, not a published survey. The rest is built from published dental-industry surveys and practitioner summaries (the ADA Health Policy Institute Survey of Dental Practice, and dental CPA and consultant surveys). "Everything else" is total overhead minus the named lines. Ranges, not targets. Last reviewed October 2026.

How the money actions are ranked. For each line above its range, the officer estimates what a full year would look like if the line moved back to the top of its range, and ranks the actions by that dollar figure. For a line already inside its range but in the upper half while total overhead runs high, the figure is half the distance to the middle of the range, and the sentence says so. Actions worth less than 1% of collections drop to the bottom. The dollar figures are rough sizes, not forecasts or promises.

Practice-software reports. When you drop in a month-end report (PDF or CSV), the office looks for the totals every system prints under similar labels: production, collections, new patients, hygiene reappointment, case acceptance, broken appointments, unscheduled treatment and patients seen. Lines with averages, goals or per-visit figures are skipped. It shows you each number it found next to the line it came from, so you can check it, and it reads the file on your computer: nothing is uploaded.

Bank files. Each bank line is sorted with rules for dental vendors and payers we know (for example Henry Schein, Patterson, Benco, Darby and Glidewell; payroll services; dental insurers and patient financing). Transfers, card payments and owner draws are left out. You can move any line; the fix is saved on your device and applies to that vendor from then on.

The AI Operations Officer

Four front-office measures have typical ranges cited in dental consultant and practice-software articles (the new-patient range is our own estimate from those articles), and the officer compares you with them:

MeasureUsual rangeNote
New patients last month20 to 40Our estimate from dental consultant and practice-software articles; wide variation by market and practice size.
Hygiene patients reappointed before leaving80 to 90%Targets in consultant and practice-software articles run 80 to 90%; many offices report less.
Case acceptance40 to 60%Averages cited in consultant and practice-software articles run 40 to 60 percent; depends on how acceptance is counted.
Collections as a share of production95 to 100%Consultant and practice-software articles cite 98 percent or better as the aim; below 95 deserves a look at your open balances.

These are typical ranges reported by dental consultants and practice-software publishers, not a national survey, and offices count them differently.

Patients seen, broken appointments (no-shows plus same-day cancellations) and unscheduled treatment have no fair outside range we could stand behind, so the officer compares each only with your own last month. Two rules of thumb of our own: broken appointments at 10% or more, or up more than two points in a month, are worth a look; and any unscheduled treatment gets a call list.

The AI Marketing Officer

It reads your own numbers: marketing spend divided by new patients, your Google rating and review count, new reviews this month, and where new patients said they came from, and it compares each with your own last month. It also uses four rules of thumb of our own, which are not published ranges:

AI-written text

Nothing in the free read is written by an AI model: every report, range and action is the software applying the rules on this page. A separate, optional AI-written note (Claude, from Anthropic) is limited to private beta testers; it is not part of what you see here, and nothing is sent to it unless a beta tester presses its button.

This month's three

Each month the office takes the top action from each officer, so you get exactly three things to do, one per seat. If an officer has nothing worth doing that month, its slot goes to the next action from the others.

The year-end checklist

The year-end checklist uses 2026 figures from IRS and California Franchise Tax Board sources, each listed with its link at the bottom of the year-end checklist.

What it is not