Skip to content
Prompt Library intermediate

How to Use AI for Dental Insurance Narratives: A Prompt Library for Common Procedures

Insurance narratives are the most tedious writing in dental practice — repetitive, formulaic, and necessary. AI handles the structure well but gets clinical details wrong in ways that could get a claim denied. Here are 8 prompt templates for common procedures, with outputs, grades, and the specific edits a practice needs to make before submitting.

Published | Updated | Verified

Insurance narratives are the dental equivalent of TPS reports. They’re necessary, formulaic, and soul-crushingly repetitive. Every crown needs a narrative justifying medical necessity. Every root canal needs documentation. Every implant needs pre-authorization with supporting clinical findings. The language is specific, the structure is predictable, and the consequences of getting it wrong are real — a badly written narrative doesn’t just get denied, it delays treatment and frustrates patients who thought their insurance would cover it.

This is where AI should excel. Insurance narratives follow a rigid pattern: clinical findings, diagnosis, proposed treatment, rationale for medical necessity. The vocabulary is standardized. The structure barely varies. It’s exactly the kind of repetitive professional writing that large language models were built for.

And it mostly works. Mostly. AI consistently produces narratives with the right structure, appropriate clinical terminology, and defensible medical necessity language. Where it fails is in clinical specifics — it will state findings that weren’t in your prompt, imply diagnoses you didn’t make, or use CDT code language that doesn’t match the actual procedure. These are the kinds of errors that get claims flagged, and they’re subtle enough that a tired office manager might miss them on a busy Friday afternoon.

Testing honesty: We tested all prompts using Claude (claude-sonnet-4-6) and ChatGPT (GPT-4o) through their web interfaces. Outputs shown are unedited. Insurance narratives were reviewed for CDT code accuracy against published ADA guidelines by a practicing dentist, not by a dental insurance specialist or claims reviewer. We did not submit any of these generated narratives to actual insurance companies. Your results will vary by carrier, plan type, and regional customs.


Prompt 1: Crown Narrative — D2740 (Porcelain/Ceramic, Molar)

The scenario: a patient needs a crown on tooth #14 (upper left first molar). There’s an existing large amalgam restoration with recurrent decay, and the mesiolingual cusp is fractured. You need a narrative for the pre-authorization.

The prompt:

You are writing an insurance pre-authorization narrative for a dental procedure.

Patient: [Patient Name], DOB [Date of Birth]
Provider: Dr. Sarah Chen, DDS
Practice: Maple Street Dental

Procedure: D2740 — Crown, porcelain/ceramic substrate
Tooth: #14 (upper left first molar)

Clinical findings:
- Existing large MOD amalgam restoration, approximately 15 years old
- Recurrent decay at the distal margin of the existing restoration, confirmed on bitewing radiograph
- Fracture of the mesiolingual cusp extending below the gingival margin
- Tooth is vital, responding normally to cold test
- Insufficient remaining tooth structure to support a direct restoration

Diagnosis: Recurrent caries with cusp fracture on previously restored tooth

Write a pre-authorization narrative supporting medical necessity for a porcelain/ceramic crown. Use standard insurance narrative structure: clinical findings, diagnosis, rationale for crown vs. direct restoration. Keep it under 200 words. Use clinical terminology appropriate for insurance reviewers.

The output:

Patient presents with tooth #14 exhibiting recurrent caries at the distal margin of an existing large MOD amalgam restoration, approximately 15 years old, as confirmed on bitewing radiographic examination. Clinical examination reveals fracture of the mesiolingual cusp extending subgingivally.

Vitality testing with cold stimulus produced a normal response, indicating the pulp is vital and endodontic therapy is not indicated at this time.

The combination of recurrent decay requiring removal of the existing restoration, loss of the mesiolingual cusp, and the extent of the original MOD preparation leaves insufficient coronal tooth structure to predictably support a direct restoration (composite or amalgam). A full-coverage crown is necessary to restore the tooth to function, protect the remaining tooth structure from further fracture, and prevent the need for more extensive treatment such as root canal therapy or extraction.

D2740 (porcelain/ceramic substrate) is selected for this molar based on the patient’s occlusal requirements and esthetic considerations. The restoration will provide adequate strength for posterior function while maintaining acceptable esthetics.

Supporting documentation: bitewing radiograph, clinical photographs, and periodontal charting are available upon request.

Grade: Ready to Submit

This is a clean narrative. The structure follows the standard insurance format: findings first, vitality status, rationale for crown over direct restoration, and code justification. The language about “insufficient coronal tooth structure to predictably support a direct restoration” is the key medical necessity phrase that insurance reviewers look for. One minor note: the AI added “periodontal charting” to the supporting documentation list, which is appropriate for comprehensive records but not strictly necessary for a crown pre-auth. The reviewing dentist confirmed every clinical statement matches the findings provided in the prompt — the AI didn’t invent any findings. This could be submitted with practice-specific formatting adjustments only (letterhead, patient demographics, provider NPI).


Prompt 2: Root Canal Narrative — D3330 (Molar Endodontic Therapy)

The scenario: tooth #19 (lower left first molar) with symptomatic irreversible pulpitis. Patient has been experiencing spontaneous pain, lingering response to cold, and the periapical radiograph shows a radiolucency at the mesial root apex.

The prompt:

You are writing an insurance pre-authorization narrative for a dental procedure.

Patient: [Patient Name], DOB [Date of Birth]
Provider: Dr. Sarah Chen, DDS
Practice: Maple Street Dental

Procedure: D3330 — Endodontic therapy, molar (excluding final restoration)
Tooth: #19 (lower left first molar)

Clinical findings:
- Patient reports spontaneous, severe pain on the lower left lasting 30+ seconds
- Lingering painful response to cold stimulus on tooth #19 (response persists after stimulus removed)
- Percussion sensitivity on tooth #19
- Periapical radiograph shows widened PDL space and a 2mm radiolucency at the mesial root apex
- Deep existing MO composite restoration present
- No mobility, probing depths within normal limits (3mm or less circumferentially)

Diagnosis: Symptomatic irreversible pulpitis with symptomatic apical periodontitis, tooth #19

Write a pre-authorization narrative supporting medical necessity for endodontic therapy. Use standard insurance narrative format. Keep it under 200 words.

The output:

Patient presents with spontaneous, severe pain localized to tooth #19, with episodes lasting greater than 30 seconds. Cold vitality testing elicits a lingering painful response that persists well beyond removal of the thermal stimulus, consistent with irreversible pulpal inflammation. The tooth is sensitive to percussion.

Radiographic examination reveals a deep existing MO composite restoration in close proximity to the pulp chamber, a widened periodontal ligament space, and a 2mm periapical radiolucency at the mesial root apex, indicative of apical pathology.

Periodontal evaluation shows probing depths within normal limits (3mm or less) circumferentially, and the tooth exhibits no mobility, confirming that the pathology is endodontic in origin rather than periodontal.

Clinical and radiographic findings are consistent with a diagnosis of symptomatic irreversible pulpitis with symptomatic apical periodontitis (AAE classification). The tooth is restorable and has a favorable long-term prognosis with endodontic therapy.

Endodontic therapy (D3330) is indicated to remove the irreversibly inflamed pulp tissue, debride the canal system, and resolve the periapical pathology. Without treatment, the condition will progress to pulp necrosis and potential abscess formation, likely resulting in the need for extraction.

A final restoration (crown) will be treatment planned separately following successful endodontic therapy.

Grade: Ready to Submit

The narrative hits every point an insurance reviewer needs: subjective symptoms, objective testing results, radiographic findings, periodontal status (ruling out a perio-endo lesion), diagnosis using AAE classification, restorability assessment, and prognosis. The addition of “AAE classification” is a strong detail — it signals to the reviewer that the diagnosis follows established endodontic diagnostic criteria. The line about progression to necrosis and extraction provides the medical necessity justification. The mention of a future crown is appropriate and shows treatment planning continuity. The reviewing dentist flagged one thing: the AI added “in close proximity to the pulp chamber” about the existing restoration, which is a reasonable inference but wasn’t explicitly stated in the prompt findings. If the restoration isn’t actually close to the pulp on the radiograph, remove that detail. Otherwise, submit with demographic formatting.


Prompt 3: Implant Narrative — D6010 (Surgical Placement, Endosteal)

The scenario: tooth #19 was extracted 4 months ago. The patient wants a single implant to replace the missing molar. The ridge is adequate for implant placement. The opposing tooth is present and functional.

The prompt:

You are writing an insurance pre-authorization narrative for a dental procedure.

Patient: [Patient Name], DOB [Date of Birth]
Provider: Dr. Sarah Chen, DDS
Practice: Maple Street Dental

Procedure: D6010 — Surgical placement of implant body, endosteal implant
Site: #19 (lower left first molar) — extracted approximately 4 months ago

Clinical findings:
- Tooth #19 extracted 4 months prior due to non-restorable fracture
- Adequate ridge height and width confirmed on CBCT scan (minimum 10mm height, 7mm width at the #19 site)
- Adequate keratinized tissue present at the site
- Opposing tooth #14 is present and fully functional
- Adjacent teeth #18 and #20 are unrestored and caries-free
- No contraindications to surgical placement (patient is non-smoker, no uncontrolled systemic conditions)

Diagnosis: Partially edentulous, missing tooth #19

Write a pre-authorization narrative supporting medical necessity for implant placement. Include rationale for implant over alternative tooth replacement options (fixed bridge, removable partial denture). Keep it under 250 words.

The output:

Patient presents with a missing tooth #19 (lower left first molar), which was extracted approximately four months ago secondary to a non-restorable vertical root fracture. The extraction site has healed uneventfully. Cone beam computed tomography (CBCT) confirms adequate residual ridge dimensions for implant placement, with a minimum of 10mm available bone height and 7mm buccolingual width at the proposed implant site. Adequate keratinized tissue is present.

The opposing tooth #14 is present and fully functional, confirming the presence of an opposing occlusal partner. Loss of tooth #19 without replacement will result in supraeruption of the opposing tooth, mesial drifting of tooth #18, and potential loss of occlusal stability in the posterior quadrant. Functional mastication on the left side is compromised.

Alternative treatment options were considered:

  • Fixed bridge (D6740/D6750): A conventional three-unit fixed bridge would require preparation of adjacent teeth #18 and #20. Both teeth are unrestored and caries-free. Preparing intact teeth for bridge abutments would result in unnecessary removal of healthy tooth structure and increased long-term risk of pulpal complications on two additional teeth.
  • Removable partial denture (D5213/D5214): A removable prosthesis provides inferior functional outcomes, reduced chewing efficiency, and lower patient acceptance compared to a fixed implant-supported restoration.

Implant placement (D6010) is the most conservative and predictable treatment option. It replaces the missing tooth without compromising adjacent dentition, restores posterior occlusal function, and prevents the adverse consequences of an unreplaced missing molar. The patient is a non-smoker with no uncontrolled systemic conditions, presenting a favorable risk profile for implant osseointegration.

CBCT scan, clinical photographs, and periapical radiographs are available upon request.

Grade: Needs Clinical Review

The structure is strong and the alternatives comparison is exactly what implant pre-auths require — insurance reviewers want to see that you considered and rejected less expensive options with clear clinical reasoning. The rationale for rejecting a fixed bridge (both adjacent teeth are unrestored) is the correct argument and is well-articulated.

However, two clinical details need verification. First, the AI described the extraction reason as “vertical root fracture.” The prompt said “non-restorable fracture” without specifying the type. If the actual fracture was a crown fracture rather than a vertical root fracture, the narrative needs correction — the distinction matters because vertical root fractures have implications for the surrounding bone and healing. Second, the AI included CDT codes for the alternative treatments (D6740/D6750 for bridge, D5213/D5214 for RPD). While this is a nice structural touch, verify these are the correct codes for your proposed alternatives. D5213 is a maxillary RPD, not mandibular — the correct code for a mandibular RPD would be D5214. The reviewing dentist caught that immediately. These are the kinds of details that look right at a glance but could trigger a question from a careful reviewer. Fix the fracture description if needed, correct the RPD code, and this is submittable.


Prompt 4: Scaling and Root Planing — D4341

The scenario: a patient with generalized moderate chronic periodontitis. Probing depths 5-7mm with bleeding on probing in multiple quadrants. The practice wants to pre-authorize four quadrants of SRP.

The prompt:

You are writing an insurance pre-authorization narrative for a dental procedure.

Patient: [Patient Name], DOB [Date of Birth]
Provider: Dr. Sarah Chen, DDS
Practice: Maple Street Dental

Procedure: D4341 — Periodontal scaling and root planing, per quadrant (4 quadrants requested)
Quadrants: UR, UL, LR, LL

Clinical findings:
- Generalized probing depths of 5-7mm with bleeding on probing in all four quadrants
- Generalized moderate horizontal bone loss confirmed on full mouth series radiographs
- Heavy subgingival calculus deposits noted on exploration
- Gingival inflammation with erythema and edema throughout
- No tooth mobility
- Patient has not had periodontal treatment previously

Diagnosis: Generalized moderate chronic periodontitis (Stage II, Grade A — per 2017 AAP/EFP classification)

Write a pre-authorization narrative supporting medical necessity for four quadrants of scaling and root planing. Keep it under 200 words.

The output:

Patient presents with generalized moderate chronic periodontitis, classified as Stage II, Grade A per the 2017 AAP/EFP Periodontal Classification System. This is the patient’s initial diagnosis; no prior periodontal treatment has been performed.

Clinical examination reveals generalized probing depths of 5-7mm across all four quadrants with bleeding on probing, indicating active inflammatory disease. Gingival tissues exhibit erythema and edema consistent with chronic inflammation. Subgingival exploration confirms heavy calculus deposits throughout.

Radiographic findings on full mouth series demonstrate generalized moderate horizontal bone loss, corroborating the clinical diagnosis. No tooth mobility is present, and the overall prognosis with appropriate intervention is favorable.

A prophylaxis (D1110) is insufficient for this patient. The presence of probing depths exceeding 4mm, subgingival calculus, bleeding on probing, and radiographic bone loss necessitate definitive periodontal therapy. Scaling and root planing (D4341) in all four quadrants is indicated to remove subgingival calculus and bacterial deposits, reduce probing depths, and arrest disease progression.

Without treatment, the condition will progress to advanced bone loss, increased probing depths, tooth mobility, and eventual tooth loss.

Re-evaluation (D4910 — periodontal maintenance) will be initiated 4-6 weeks following completion of active therapy to assess treatment response.

Grade: Needs Clinical Review

The narrative structure is solid, and the explicit statement that “a prophylaxis (D1110) is insufficient” is the right move — this is the specific justification insurance reviewers need to approve SRP over a regular cleaning. The 2017 AAP/EFP classification reference adds credibility.

Two issues the reviewing dentist flagged. First, the AI stated D4910 (periodontal maintenance) would follow at 4-6 weeks. D4910 is the correct code for periodontal maintenance visits, but the re-evaluation visit immediately following SRP is typically coded as D0180 (comprehensive periodontal evaluation) — D4910 applies to ongoing maintenance visits after active therapy is complete. This code distinction matters because some carriers will deny a D4910 billed too soon after SRP completion. Second, the AI described the periodontitis as “chronic” — the current 2017 AAP/EFP classification actually dropped the “chronic” and “aggressive” subtypes in favor of the staging and grading system. Saying “Stage II, Grade A” is correct; adding “chronic” is technically outdated terminology. Most insurance reviewers won’t flag this because many are still using older classification language themselves, but if you want the narrative to be technically precise, drop “chronic.” Otherwise the clinical findings, treatment rationale, and medical necessity language are all appropriate.


Prompt 5: Crown Buildup — D2950

The scenario: the patient from Prompt 1 needs a buildup before the crown because after caries removal and old amalgam removal, there isn’t enough tooth structure to retain the crown.

The prompt:

You are writing an insurance narrative for a dental procedure. This narrative accompanies a crown pre-authorization (D2740 on tooth #14, submitted separately).

Patient: [Patient Name], DOB [Date of Birth]
Provider: Dr. Sarah Chen, DDS
Practice: Maple Street Dental

Procedure: D2950 — Core buildup, including any pins when required
Tooth: #14 (upper left first molar)

Clinical findings:
- After removal of the existing large MOD amalgam and excavation of recurrent decay, less than 50% of the coronal tooth structure remains
- The mesiolingual cusp is fractured and has been removed
- Two walls of the preparation remain intact (buccal and palatal); the mesial and distal walls are significantly compromised
- The tooth requires a core buildup to establish adequate retention and resistance form for the crown

Write a brief narrative (under 120 words) justifying medical necessity of the buildup as a separate procedure from the crown. Insurance companies frequently deny D2950 as being included in the crown preparation — the narrative must specifically address why the buildup is a distinct procedure in this case.

The output:

Following removal of the existing MOD amalgam restoration and excavation of recurrent caries on tooth #14, clinical evaluation reveals that less than 50% of the original coronal tooth structure remains. The mesiolingual cusp has been lost to fracture. Only the buccal and palatal walls are structurally intact; the mesial and distal walls are significantly undermined and cannot independently provide adequate retention or resistance form for crown placement.

A core buildup (D2950) is required as a separate restorative procedure to replace the missing tooth structure and re-establish a preparation with adequate dimensions, retention, and resistance form to support the crown restoration (D2740). This buildup replaces lost tooth structure — it is not part of the crown preparation itself, which involves reduction of existing tooth structure. The two procedures are clinically distinct.

Grade: Ready to Submit

This narrative does the one thing it absolutely must do: it distinguishes the buildup from the crown preparation. The closing sentence — “This buildup replaces lost tooth structure; it is not part of the crown preparation itself, which involves reduction of existing tooth structure” — is the exact argument insurance companies need to hear. Buildups are one of the most commonly denied procedures because carriers argue they’re included in the crown prep fee, and most narratives don’t explicitly address the distinction. The “less than 50% coronal tooth structure remaining” is the threshold many carriers use as the standard for buildup necessity. The reviewing dentist approved this without changes. Attach it to the D2740 narrative and submit together.


Prompt 6: Dental Claim Appeal — Denied Crown

The scenario: the crown pre-authorization from Prompt 1 was denied. The insurance company’s reason: “insufficient documentation to support medical necessity.” The practice needs to write an appeal letter.

The prompt:

You are writing a dental insurance appeal letter. A pre-authorization for a crown was denied.

Patient: [Patient Name], DOB [Date of Birth], Member ID: [Member ID]
Provider: Dr. Sarah Chen, DDS, NPI: [NPI Number]
Practice: Maple Street Dental

Original claim:
- Procedure: D2740 — Crown, porcelain/ceramic substrate
- Tooth: #14 (upper left first molar)
- Date of pre-authorization submission: [Date]
- Denial reason stated by carrier: "Insufficient documentation to support medical necessity for the requested procedure"

Clinical findings (same as original submission):
- Existing large MOD amalgam restoration, approximately 15 years old
- Recurrent decay at the distal margin confirmed on bitewing radiograph
- Fracture of the mesiolingual cusp extending below the gingival margin
- Tooth is vital
- Insufficient remaining tooth structure to support a direct restoration

Additional documentation being included with the appeal:
- Bitewing radiograph showing recurrent decay
- Intraoral photograph showing cusp fracture
- Periapical radiograph of tooth #14
- Copy of the original narrative

Write a formal appeal letter. Address it to the dental claims review department. Request a re-review with the additional documentation. Be firm but professional — do not be adversarial. Keep it under 300 words.

The output:

[Practice Letterhead]

[Date]

Dental Claims Review Department [Insurance Carrier Name] [Carrier Address]

Re: Appeal of Denied Pre-Authorization Patient: [Patient Name] DOB: [Date of Birth] Member ID: [Member ID] Procedure: D2740 — Crown, porcelain/ceramic substrate, Tooth #14 Provider: Dr. Sarah Chen, DDS | NPI: [NPI Number]

Dear Claims Review Department,

I am writing to appeal the denial of the above-referenced pre-authorization. The stated reason for denial was “insufficient documentation to support medical necessity for the requested procedure.” I respectfully disagree with this determination and am submitting additional documentation for re-review.

Tooth #14 presents with an existing large MOD amalgam restoration placed approximately 15 years ago. Current bitewing radiography confirms recurrent caries at the distal margin of this restoration. Clinical examination reveals a fracture of the mesiolingual cusp extending below the gingival margin. The tooth tested vital.

Following removal of the existing restoration and carious tooth structure, the remaining coronal structure is insufficient to support a predictable direct restoration. The combination of the large existing preparation, active recurrent decay requiring further excavation, and structural loss from the cusp fracture makes a full-coverage crown the appropriate standard of care for this tooth.

Enclosed with this appeal:

  1. Bitewing radiograph demonstrating recurrent caries at the distal margin
  2. Intraoral photograph documenting the mesiolingual cusp fracture
  3. Periapical radiograph of tooth #14
  4. Copy of the original pre-authorization narrative

I request that a licensed dental consultant review this appeal with the supporting documentation. If additional information is needed, please contact our office directly at [Phone Number].

Thank you for your prompt attention to this matter.

Respectfully,

Dr. Sarah Chen, DDS Maple Street Dental [Address] [Phone Number] NPI: [NPI Number]

Grade: Ready to Submit

This is where AI is at its strongest. Appeal letters are structured argumentation — restate the facts, identify the gap in the reviewer’s evaluation, provide the evidence, and request reconsideration. The AI produced a letter with the right tone (firm but not adversarial), the right structure (reference numbers up front, clinical summary, documentation list, specific request for review by a licensed consultant), and appropriate clinical language. The request for review by “a licensed dental consultant” is a smart inclusion — it signals that the practice expects a clinician, not just an administrative reviewer, to evaluate the appeal. The reviewing dentist said she would send this with no clinical changes. The only edits needed are filling in the carrier name, address, and specific dates. This prompt alone probably saves most practices 30-45 minutes of frustrated writing per denied claim.


Prompt 7: Pre-Authorization for Orthodontics (Limited, Adult) — D8010

The scenario: an adult patient wants limited orthodontic treatment for anterior crowding. The practice offers Invisalign and needs to submit a pre-authorization to the patient’s dental plan, which has an orthodontic benefit.

The prompt:

You are writing an insurance pre-authorization narrative for orthodontic treatment.

Patient: [Patient Name], DOB [Date of Birth] (adult patient, age 34)
Provider: Dr. Sarah Chen, DDS (Invisalign-certified provider)
Practice: Maple Street Dental

Procedure: D8010 — Limited orthodontic treatment of the adult dentition
Treatment modality: Clear aligner therapy (Invisalign)

Clinical findings:
- Moderate crowding of the lower anterior teeth (4mm of arch length discrepancy)
- Mild crowding of the upper anterior teeth (2mm of arch length discrepancy)
- Class I molar relationship bilaterally
- Normal overjet and overbite
- No TMJ symptoms
- Periodontal status: healthy, probing depths 1-3mm throughout
- Estimated treatment duration: 8-12 months
- Estimated number of aligners: 18-24

Write a pre-authorization narrative supporting medical necessity for limited orthodontic treatment. Insurance carriers are skeptical of adult ortho cases — the narrative needs to address functional concerns beyond esthetics. Keep it under 200 words.

The output:

Patient is a 34-year-old adult presenting with moderate mandibular anterior crowding (4mm arch length discrepancy) and mild maxillary anterior crowding (2mm arch length discrepancy). Molar relationship is Class I bilaterally. Overjet and overbite are within normal limits. Temporomandibular evaluation is unremarkable. Periodontal status is healthy with probing depths of 1-3mm throughout.

The mandibular anterior crowding creates overlapping contact points that compromise the patient’s ability to maintain adequate oral hygiene in the lower anterior region, increasing the risk of interproximal caries and localized periodontal disease. Alignment of the lower anterior teeth will improve access for brushing and flossing, reducing long-term risk of dental disease in this region.

Limited orthodontic treatment (D8010) using clear aligner therapy is indicated to resolve the crowding and establish proper alignment. Estimated treatment duration is 8-12 months with approximately 18-24 aligners. Treatment is limited to alignment correction and does not involve molar movement or significant anteroposterior correction.

This case is appropriately classified as limited treatment (D8010) rather than comprehensive treatment (D8020) based on the scope of tooth movement required and the absence of skeletal discrepancy.

Grade: Not Submittable

The structure looks right. The language sounds professional. And this narrative will almost certainly be denied.

The fundamental problem: the clinical findings don’t support a strong medical necessity argument, and the AI couldn’t manufacture one from the data provided. The narrative’s functional justification — that crowding “creates overlapping contact points that compromise the patient’s ability to maintain adequate oral hygiene” — is technically true but weak. Insurance reviewers hear this argument constantly and most carriers don’t consider hygiene difficulty alone as sufficient medical necessity for orthodontic treatment in adults.

What’s missing is everything the AI can’t fabricate: cephalometric measurements (SNA, SNB, ANB angles), photographs showing the actual severity of the crowding, a clear statement about functional impairment beyond hygiene (is there traumatic occlusion? accelerated wear on specific teeth? food impaction causing recurring localized periodontitis?). Most orthodontic pre-authorizations for adults require specific diagnostic records — panoramic radiograph, cephalometric radiograph, study models or digital impressions, intraoral and extraoral photographs. The narrative alone, no matter how well-written, isn’t sufficient.

The reviewing dentist was direct: “This reads like it was written by someone who knows the vocabulary but hasn’t submitted many ortho pre-auths. The carrier is going to come back and ask for records that the narrative doesn’t even reference. For adult ortho, you essentially have to build the case that not treating will result in progressive deterioration — and 4mm of lower anterior crowding in a patient with healthy perio doesn’t clear that bar on its own.”

If you’re submitting adult ortho pre-auths, the AI can help structure the narrative after you’ve assembled the diagnostic records and identified the specific functional concerns. But it can’t build the clinical case for you. This is a prompt where the human expertise has to come first.


Prompt 8: Comprehensive Narrative for Multiple Procedures

The scenario: a patient needs scaling and root planing in all four quadrants, followed by crowns on teeth #3 and #14, and a possible root canal on #3 depending on post-SRP evaluation. The practice wants a single comprehensive narrative that ties the treatment plan together.

The prompt:

You are writing a comprehensive insurance narrative for a patient who requires multiple procedures.

Patient: [Patient Name], DOB [Date of Birth]
Provider: Dr. Sarah Chen, DDS
Practice: Maple Street Dental

Treatment plan (phased):

Phase 1 — Periodontal therapy:
- D4341 x4 quadrants — Scaling and root planing
- Clinical findings: generalized 4-6mm probing depths, bleeding on probing, moderate subgingival calculus, generalized mild to moderate horizontal bone loss on FMX

Phase 2 — Restorative (to be initiated 6-8 weeks after completion of Phase 1, pending periodontal re-evaluation):
- D2740 on tooth #14 — Crown, porcelain/ceramic substrate
  - Findings: large existing MOD amalgam with recurrent decay, fractured ML cusp
- D2740 on tooth #3 — Crown, porcelain/ceramic substrate
  - Findings: large existing DO composite with recurrent decay, tooth symptomatic to cold (lingering response — 5 seconds)
- D3330 on tooth #3 — Root canal, molar (may be necessary based on clinical response; will be determined at Phase 2 evaluation)
  - Current findings: lingering cold response, no periapical radiolucency currently visible

Write a comprehensive narrative that presents the full treatment plan to the insurance carrier, explains the phased approach, and justifies each procedure. The phasing is clinically important — doing restorative work before periodontal therapy is resolved would be inappropriate. Keep it under 350 words.

The output:

Patient presents with concurrent periodontal disease and restorative needs requiring a phased treatment approach. The treatment plan addresses active periodontal disease first, followed by definitive restorative treatment once periodontal stability is established.

Phase 1 — Periodontal Therapy (D4341, four quadrants)

Clinical examination reveals generalized probing depths of 4-6mm with bleeding on probing across all four quadrants. Subgingival exploration confirms moderate calculus deposits. Full mouth series radiographs demonstrate generalized mild to moderate horizontal bone loss. A prophylaxis is insufficient given the presence of probing depths exceeding 4mm, active bleeding, subgingival calculus, and radiographic bone loss. Scaling and root planing in all four quadrants is indicated to reduce bacterial load, resolve gingival inflammation, and establish a stable periodontal foundation prior to restorative treatment.

Periodontal re-evaluation will be performed 6-8 weeks following completion of Phase 1 to assess tissue response before initiating restorative procedures.

Phase 2 — Restorative Treatment (pending periodontal re-evaluation)

Tooth #14 (D2740 — Crown): Existing large MOD amalgam restoration with recurrent caries at the margin and fracture of the mesiolingual cusp. Insufficient remaining tooth structure for a direct restoration. Full-coverage crown is indicated.

Tooth #3 (D2740 — Crown, with possible D3330 — Endodontic therapy): Existing large DO composite restoration with recurrent decay. The tooth is currently symptomatic to cold testing with a lingering response of approximately 5 seconds, which is concerning for early irreversible pulpitis. No periapical radiolucency is currently visible on radiographs. Dr. Chen will re-evaluate the pulpal status of tooth #3 at the Phase 2 appointment. If symptoms have progressed or the tooth demonstrates signs of irreversible pulpitis at that time, endodontic therapy (D3330) will be initiated prior to crown fabrication. If symptoms have resolved following periodontal therapy and the tooth tests within normal limits, the crown will proceed without endodontic intervention.

Rationale for Phased Approach

Performing definitive restorative treatment in the presence of active periodontal disease compromises crown margin integrity, sulcular health around the restorations, and long-term prognosis of the restored teeth. Periodontal stability must be established before investing in full-coverage restorations.

Grade: Needs Clinical Review

This is a well-organized narrative that does something most insurance submissions fail to do: it explains why the treatment is phased and why that phasing matters clinically. The argument that restorative treatment in the presence of active periodontal disease “compromises crown margin integrity” is legitimate and gives the insurance reviewer a clear reason to approve the sequencing.

The tooth #3 section is particularly well-handled — the AI correctly presented the root canal as conditional, explained the current clinical findings that raise concern, and described the decision point at the Phase 2 evaluation. This is exactly how a clinician would present an uncertain endodontic prognosis.

Two items to verify. First, the AI described the cold response on tooth #3 as “approximately 5 seconds” — the prompt said “lingering response — 5 seconds.” Five seconds of lingering is borderline. Some clinicians would consider 5 seconds still within normal limits for a heavily restored tooth; others would call it early irreversible pulpitis. Make sure the 5-second characterization matches your actual clinical finding and your diagnostic interpretation. Second, the narrative doesn’t mention the periodontal classification (staging and grading). Adding “Stage II, Grade A” or whatever the appropriate classification is would strengthen the periodontal section. The reviewing dentist recommended adding it. Otherwise, the structure and logic of this narrative are sound. Review the clinical details against the actual chart and submit.


What AI Gets Right About Insurance Narratives

The structure is consistently good. Every narrative the AI produced followed the standard format that insurance reviewers expect: clinical findings, diagnosis, proposed treatment, rationale for medical necessity. It didn’t need to be told this format — it learned it from the millions of dental narratives in its training data. That alone saves time, because the most tedious part of writing narratives isn’t the clinical content, it’s building the framework around it.

The medical necessity language is appropriate. Phrases like “insufficient coronal tooth structure to predictably support a direct restoration,” “a prophylaxis is insufficient given the presence of probing depths exceeding 4mm,” and “restores posterior occlusal function and prevents adverse consequences” are exactly what insurance reviewers look for. The AI consistently used the vocabulary that signals clinical justification rather than just treatment description. This is the difference between a narrative that gets approved and one that gets sent back for more information.

The appeal letter was the strongest output. This makes sense — appeal letters are pure argumentation, and AI handles structured argumentation well. The combination of restating facts, identifying the gap, providing evidence, and requesting specific action follows a logical pattern that AI excels at. If you’re currently writing appeal letters from scratch, this is the single highest-value prompt in this library.

What AI Gets Wrong

Clinical specifics are the consistent failure point. The AI added “in close proximity to the pulp chamber” to the root canal narrative when that finding wasn’t in the prompt. It described a fracture as “vertical root fracture” when the prompt said “non-restorable fracture.” It used the wrong RPD code (D5213 instead of D5214) in the implant alternatives section. It described periodontitis as “chronic” using outdated classification terminology. Each of these is a small error. Each could trigger a question from a reviewer, slow down the approval process, or worse, create a discrepancy between the narrative and the clinical record.

The orthodontic narrative exposed the deepest limitation. When the clinical case is weak, the AI can’t strengthen it. It can polish the language and organize the presentation, but it can’t manufacture clinical findings, produce diagnostic records, or identify functional justifications that aren’t in the prompt. For straightforward procedures where the clinical indications are clear — crowns, root canals, SRP — the AI drafts are usable. For cases where the medical necessity argument requires clinical judgment and experience to construct, the AI is starting from a deficit.

The Right Workflow

The workflow that works for insurance narratives is three steps, and none of them are optional.

Step one: the clinician provides the clinical findings. Not the AI. The dentist or the hygienist documents what they found — probing depths, radiographic findings, vitality test results, specific tooth structure remaining, specific clinical observations. If your practice management system has structured charting (and if you are on Dentrix Ascend, this is built into the platform), export those findings directly rather than retyping them. This is the raw material the narrative is built from, and it must come from the person who examined the patient.

Step two: the AI drafts the narrative from those findings. Use the prompts above, fill in the clinical details, and generate the draft. This takes 30 seconds and produces a structurally sound narrative with appropriate medical necessity language.

Step three: the clinician reviews every clinical statement in the draft before it goes out the door. Not the office manager. Not the insurance coordinator. The person who examined the patient and made the diagnosis reads the narrative and confirms that every finding stated in it matches the chart. This takes two minutes. It catches the fabricated details, the wrong codes, the outdated terminology, and the clinical inferences the AI made that don’t match reality.

Skip step three and you’re submitting narratives with clinical statements that may not match your records. That’s worse than a poorly written narrative — it’s a discrepancy between your documentation and your insurance submission, and that’s the kind of thing that creates problems beyond a simple denial.

The AI saves time on the writing. It does not replace the clinical judgment. Keep those two things separate and insurance narratives become a five-minute task instead of a twenty-minute task, without increasing your risk. We found the same boundary when we tested Claude, ChatGPT, and Gemini on treatment plan explanations — structure was strong across all three models, but clinical details needed human review every time.

Sources