Before the visit
An appointment from the practice calendar or an authorized Doctolib connection triggers an intake form. Doctolib is a widely used French booking platform. The practice reviews medications, allergies and the reason for the visit.
Investor brief · November 2026
The exam
becomes the
chart.
Dentists already explain their findings to patients. CodeN is building a $79/month agent that listens with permission, drafts the chart and prepares the follow-up. The dentist reviews it.
See the tested case ↘
CodeNVoice-first dental practice software · France first01 / 19
01 / Legacy software vs. CodeN
Click through forms. Or review a draft.
Same finding: decay on a lower molar. Here is how the work differs.
Legacy dental software
CodeN’s voice-first workflow
An incumbent product dates to 1983. Left: English reconstruction of a 2019 competitor manual, not a screenshot of its current release. Right: CodeN concept; the text-to-chart workflow has been tested.
02 / Inside the legacy workflow
The tooth is charted. The paperwork is still open.
The tooth chart records where the problem is. It doesn’t, by itself, describe what the patient felt, price a procedure or explain the next step to the patient.
CodeN’s proposed shortcut: use the reviewed encounter as the source for the chart, the note and the follow-up draft. The dentist checks the result instead of entering the same facts again.
English reconstructions of documented manual workflows, with illustrative data. Sources: procedure setup guide, p. 18 and practice software manual, pp. 59–60 and 156. Templates, attachments and automation vary by product and version.
03 / Inside CodeN
A closer look at the chairside workspace.
Follow one fictional exam from the full dental chart to the selected tooth and the note ready for review.
The consultation
Start with the patient, both arches and the clinical summary in one view.
The tooth inspector
Select the surfaces, record the finding and check the wording that supports it.
The exam review
Read the narrative beside the proposed chart changes before confirming the update.
English design previews using fictional data. The tooth chart and anatomical illustrations are rendered from CodeN’s own components; the surrounding screens illustrate the intended workflow.
04 / The product
One conversation can drive a whole visit.
With the patient’s permission, the dentist talks to the patient as usual. CodeN’s goal is to turn that exchange into reviewable work while the details are still fresh.
In the tested case“Occlusodistal” on tooth 36 became two charted surfaces on one tooth. “Sensitive to cold” stayed in the note as a patient-reported symptom. A planned X-ray stayed a plan.
- Listen in real time. The planned ambient mode captures speech while the microphone is on. Pause or stop at any time; use typing or an individual voice command instead.
- Read just what’s relevant. The agent resolves the patient from the open chart or a unique match, then requests the records it needs. If two patients fit, it asks one question.
- Choose the next tool as it goes. Qwen can request a patient lookup, read a chart or propose an update. Each tool returns a result that determines the next call; there is no fixed sequence of forms.
- Save one reviewed clinical update. The server checks access, source text and record version. In the prototype, approval saves the full note and all localized findings together, without duplicates on retry.
- Prepare what comes next. The intended workflow drafts the estimate, prescription, patient message or stock request that the visit calls for. Sending, signing and payer submission stay subject to approval and available integrations.
Demonstrated: French text to a reviewed note and tooth chart using the real model. Pilot questions: Can ambient capture recover the findings dentists naturally say? Our 95% spoken-coverage estimate and the time saved after review have not yet been measured in practice.
05 / The intended patient journey
One visit, from booking to checkout.
One record carries the reason for the visit, the medical history, the exam and the next steps. The goal is to finish more of the administrative work while the dentist is still with the patient.
During the exam
After the patient agrees, the dentist turns on the microphone. The planned ambient workflow listens for speech, separates speakers and lets the dentist pause or stop. If the patient declines, manual charting remains available.
At checkout
CodeN prepares the note, estimate, prescription draft and follow-up tasks from what the dentist actually decided. The dentist can add unspoken findings and approve or discard each item.
The documented clinical test started with natural-language text. Ambient audio capture during patient visits and a live Doctolib connection have not been validated.
06 / Practice operations
The agent drafts. The dentist authorizes.
A draft, a signed document, a submitted claim and a payment are different events. CodeN is designed to track which one actually happened.
WorkflowAgent’s workApproval and destination
Clinical chartKeep the full note and map explicit findings to teeth and surfaces.The dentist reviews one chart update; its source stays linked to the record.
Treatment and estimatesPrepare the proposed treatment, estimate and plain-language explanation.The dentist approves the version to print or send to the patient.
Prescriptions and follow-upDraft the prescription as directed, aftercare instructions, referral and recall.The prescriber checks and signs. Approved documents go to the patient or referring clinician through an appropriate channel.
Insurance billingMatch completed procedures to the invoice. The internal engine simulates claim submission, responses and payment reconciliation.Target: submit approved claims to the public payer and, where applicable, supplemental insurers through authorized connections.
Supplies and purchasingConnect usage, inventory thresholds and reorder requests.Send approved orders through supported supplier integrations; use an export or manual order where no API is available.
The claims engine operates in simulation. Live payer submission, electronic signatures, patient delivery and supplier orders require production integrations.
07 / Product evidence · October 27, 2026
One exam, one verified chart update.
Tooth 36 has a cavitated lesion on the occlusal and distal surfaces. The patient reports sensitivity to cold.
English translation of an excerpt from the fictional French exam submitted to Qwen. FDI 36 is the lower left first molar (U.S. tooth #19).
Recorded after test confirmation
36Decay mapped to the chewing surface and the side facing away from the midline.
46Missing tooth; no cause inferred.
31/32
41/42Tartar mapped to the tongue-facing surfaces of the four lower incisors.
NoteThe full narrative retained the symptoms, negative findings and planned X-rays.
1atomic chart update; retry testing found no duplicates
3model calls in the successful run
41 secfor that run, including initialization
Internal Windows qualification: 263 targeted tests passed. The exam used the real model, synthetic data and a separate test confirmation. No clinical efficacy study or in-office latency result is claimed.
08 / French market size
An estimated $196M in annual software spending.
We size the existing market using estimated recurring spend on the incumbent software stack. CodeN’s own subscription price does not enter this calculation.
$195.7M
Estimated national annual run rate: 48,690 active dentists × about $335 per month × 12. Equivalent to $16.31M per month.
French health statistics agency · 2026 workforce$161.1M
Estimated annual run rate for the initial private-practice segment: 40,083 private or mixed-practice dentists × about $335 × 12. Equivalent to $13.43M per month.
French provider registry · practice settingWhat the estimate means
About $335 is our working estimate of combined incumbent subscriptions per dentist, not an audited national average or a vendor revenue total. Shared licenses and group purchasing can lower the per-dentist figure. At roughly $223–$447 per month, the national estimate ranges from $130.5M to $260.9M annually.
For scale, French private practices delivered $14.7B in dental care in 2024. Clinical spending is separate from software spending. French figures are expressed in U.S. dollars at the November 6, 2026 ECB rate of US$1.1165 per euro; calculations precede rounding.
09 / Worldwide and U.S. market size
The U.S. accounts for about 42% of a $2.36B global market.
Dental practice management software covers the core systems practices use for scheduling, patient records and billing. In 2025, estimated U.S. revenue reached $990M, leaving roughly $1.37B across the rest of the world.
| Market | 2025 estimate | 2026 estimate | 2031 forecast |
|---|---|---|---|
| Worldwide | $2.36B | $2.62B | $4.44B |
| United States | $0.99B | $1.09B | $1.69B |
Why the U.S. matters
The American Dental Association counted 205,088 dentists working in dentistry in 2025. After validating the French product, CodeN could test its clinical workflow with U.S. practices before investing in local claims integration. The forecast is for the category, not CodeN sales.
How to read the figures. The worldwide and U.S. estimates come from Mordor Intelligence. The U.S. is included in the global total. These are market revenue estimates, not audited subscription ARR. France’s preceding estimate models a broader bundle of tools and should not be added to or divided by these figures. Other 2025 global estimates include $1.96B (Fortune Business Insights) and $2.63B (Global Market Insights).
10 / Competitive landscape
Two products account for 66% of French dental claim filers.
Replacing a stack, not just a chart
A practice may use LOGOSw or Julie for its core record, La Fraise for estimate follow-up, and a separate inventory system. Some practices already bundle several functions in one license.
Incumbents also offer AI. The test for CodeN is whether one subscription can replace overlapping products and reliably turn a reviewed finding into work completed in the practice system.
GIE SESAM-Vitale · September 2026, dentists, by software product. Denominator: 39,572 filers. This measures electronic billing use, not share of revenue or number of practices. More than one product can belong to the same company.
11 / What the dentist pays
About $335 for the stack. $79 for CodeN.
~$335
Current software stack / month
One core system such as LOGOSw or Julie, plus any separate subscriptions for inventory, estimates and patient follow-up. Estimated combined spend per dentist.
$79
CodeN / dentist / month
One planned subscription for charting, the voice agent, estimates, patient follow-up, billing workflows, inventory and purchasing. $948 per year.
What would your practice save?
Count only subscriptions the practice can cancel, on the same tax basis. Exclude any retained third-party services.
$256potential monthly savings per dentist
$3,072potential annual savings · about 76% lower
Price checks: practice software about $76/month per dentist; inventory about $32/month per practice; La Fraise estimates and follow-up priced by quote. The $335 benchmark is a working estimate, not a measured average. Savings use the rounded $335 comparison.
12 / Unit economics · Per dentist per month
At $79, the base case spends $12 on Qwen.
One successful synthetic exam used 47,545 input tokens and 2,246 output tokens across three calls. At uncached API list prices, that is about four cents per exam.
Assumed monthly usageLightBase caseHeavy
Exams per day × 20 days8 × 2015 × 2025 × 20
Qwen text cost$6.42$12.04$20.06
Speech-to-text at 8 spoken minutes per exam$4.29$8.04$13.40
Revenue left after text + speech$68.29$58.92$45.54
What it takes to earn a 65% gross margin
In the base case, 74.6% of revenue remains after text and speech. To reach a 65% gross margin, all other direct costs must fit within $7.57 per dentist per month, including hosting, support, transaction fees and integration costs.
Illustrative usage, not a customer average. Rates: about $0.67 per million input tokens, $3.68 per million output tokens and $0.00335 per audio minute. No free tier or caching assumed. Additional actions, retries and longer visits raise costs. Private GPU hosting requires a separate utilization model. Scaleway list prices; calculations use unrounded rates.
13 / Architecture and health data
Only reviewed findings enter the chart.
Capture
Patient permission first
The intended experience includes a visible recording indicator, speech detection, pause and stop controls. The dentist can always enter information manually.
Interpretation
A dental agent using Qwen
The model receives only relevant record data and authorized tools. It can resolve an unambiguous reference, ask about a material uncertainty and leave unknown facts unrecorded.
Execution
Checks before a write
The application checks patient identity, access rights, source and record version. In the clinical test, confirmation saved the narrative and structured findings together.
Deployment in France
The prototype called Qwen through Scaleway’s API on synthetic data. The intended clinical deployment runs Qwen privately near the records within a French HDS-certified hosting environment. HDS is France’s health data hosting certification; CodeN does not claim a certified production deployment today.
Deployment work includes legal basis, data retention, security review and a privacy impact assessment where required. French Digital Health Agency: HDS / French health and privacy authorities: AI in care.
14 / International expansion
One clinical core. A billing layer for each country.
Examining teeth, recording findings and explaining a treatment plan are recognizable across markets. Who pays, which codes are accepted and how a claim gets approved are local questions.
Start in France
France is the reimbursement system the team knows best. The product is designed to connect practices to national public health insurance and supplemental insurers, including mutual insurers. Approved claims, payer responses and remittances must reconcile with the patient ledger.
Reuse the clinical core
Carry the exam narrative, tooth-and-surface mapping, review controls and task history into the next market. Localize language, terminology, numbering, prescribing rules and consent before clinical use.
Build each payer connection
For every new market, map eligibility, procedure codes, fees and claims formats; secure payer or clearinghouse access and test rejection and payment flows. In the U.S., examples include the 837D dental claim and 835 remittance.
The goal is rapid expansion after French validation. Launch in each country depends on local approvals and successful end-to-end payer testing. Live public or supplemental payer connectivity is not claimed. France: billing software approval / U.S.: CMS dental claims guide.
15 / Proposed corporate structure
Invest in one company. Reach multiple markets.
U.S. PARENT COMPANY
CodeN · Delaware C corporation
Investors would buy shares here. The parent would raise capital, own CodeN’s proprietary software and brand, and license the platform to its operating subsidiaries.
FIRST MARKET
France · Wholly owned SASU
The French subsidiary would sign customer contracts, sell and support CodeN, and adapt billing to French public and supplemental insurers.
FUTURE MARKETS
Local operating subsidiaries
Each subsidiary would distribute the same software under license and handle local contracts, payer integrations and country-specific compliance work.
Why this structure? A Delaware corporation can issue the preferred stock familiar to U.S. venture funds. A French SASU can contract in its own name while the U.S. parent remains its sole shareholder.
Implementation requires legal and tax review of IP assignments, tax residence and intercompany license terms. Each entity retains the obligations that apply to its role. U.S. venture financing · French SASU.
16 / Investment and go-to-market
The next milestone: a paid pilot in France.
Investment would fund the move from a tested Windows prototype to daily chairside use, starting with independently owned French practices where the buyer is also a user.
Measure the baseline
Recruit pilot sites through chairside demos and professional referrals. Review their software invoices and identify subscriptions they could cancel. Time the same visit types before and after adoption.
Test ambient capture
Measure what dentists say without prompting, what the agent misses, corrections required, and net time saved after review. Track patient-facing time separately. Target under 30 seconds for 95% of simple actions.
Prove the economics
At $79 per month, 1,000 paying dentists would generate $79,000 in monthly recurring revenue, or $948,000 annually. Track conversion, retention and actual delivery cost before scaling. Expand payer and supplier connections only as partners and technical requirements are verified.
The pilot succeeds if dentists can spend more of each visit with the patient, review an accurate chart and retire software subscriptions without shifting work to the front desk.
Investor diligence: a live demonstration using a fictional case of your choice, followed by a review of the pilot protocol, unit economics and payer integration milestones.
17 / References
Sources and assumptions.
French market and competition
French dentist workforce, 2026French provider registry: private and mixed practiceFrench national health accounts: dental spendingElectronic claim filers by product, September 2026Historical procedure-entry guide, pages 2 and 18AI features in incumbent practice softwarePricing, hosting and payer infrastructure
Practice management subscription pricingInventory and ordering subscription pricingEstimate management pricing policyFrench claims software approval processFrench health data hosting certificationClinical AI and privacy guidanceBooking platform connector termsU.S. CMS: dental claims and remittance standardsInternal evidence
Windows qualification, October 27, 2026: 263 targeted tests and one successful live-model exam on synthetic data. The images are English reconstructions. The 95% speech coverage estimate, $335 incumbent budget and savings comparison are assumptions to verify in the pilot.
Cost sources: Scaleway Qwen and speech API pricing · ECB exchange-rate series. Incumbent product history: Julie corporate timeline.
18 / From the founder
I’m building CodeN with two dentists who practice in France.
My background
I started in sales and went on to serve as a sales director. I also founded an AI software-as-a-service company. With CodeN, I’m bringing those two parts of my career together: building the software and understanding why a practice would switch to it.
Our clinical collaborators
Corentin and Paul are practicing dentists in France with different specialties. Both are involved in developing CodeN. Their two perspectives help us ask a practical question: does this give dentists more time with the patient, or just another screen to manage?
I’d welcome the chance to show you the product. Bring a fictional case, and we’ll walk through what CodeN can do today and what the pilot is designed to prove.
CodeNFounder’s note · November 2026