AI experts sharing free tutorials to accelerate your business.
Back to HVAC toolkit

Visual Inspection Report

Transform technician field photos and accompanying notes into a structured visual inspection report. The AI identifies components from photos or descriptions, flags concerns, decodes nameplate data (including serial-number age decoding for the major OEMs), and produces a professional report suitable for customer delivery, warranty documentation, realtor pre-purchase packs, or internal QC review.

Saves ~20 min/reportintermediate Claude ยท ChatGPT ยท Gemini

๐Ÿ“ธ Visual Inspection Report

Purpose

Transform technician field photos and accompanying notes into a structured visual inspection report. The AI identifies components from photos or descriptions, flags concerns, decodes nameplate data (including serial-number age decoding for the major OEMs), and produces a professional report suitable for customer delivery, warranty documentation, realtor pre-purchase packs, or internal QC review.

When to Use

  • After a site visit where the tech took photos of equipment, ductwork, or installations
  • When documenting the condition of aging or damaged equipment for a replacement proposal
  • During pre-purchase home inspections covering HVAC systems (realtor-facing output)
  • When a property manager requests a written condition assessment of multiple units
  • For warranty claim support where visual evidence of a defect must be documented
  • When compiling before/after documentation for completed installations
  • When uploading structured condition data into the dispatch/CRM system for a recurring commercial account

Required Input

Provide the following:

  1. Photos or photo descriptions โ€” Upload images directly (if using a multimodal AI) or describe what each photo shows, including visible equipment labels, damage, corrosion, leaks, wiring condition, etc.
  2. Equipment info (if available) โ€” Make, model, serial number, age, system type
  3. Inspection type โ€” Routine maintenance, pre-purchase, warranty claim, post-install QC, damage assessment, or code compliance
  4. Location / customer (optional) โ€” For report header
  5. Specific concerns (optional) โ€” Any areas the customer or tech wants highlighted
  6. Audience โ€” homeowner, realtor-pre-purchase, warranty-file, property-manager, internal-qc (changes tone and disclaimer language)

Instructions

You are a senior HVAC inspection specialist producing a visual condition assessment report. Organize photo-based observations into a professional report that clearly communicates equipment condition, decodes nameplate data, and routes appropriate actions.

Before you start:

  • Load config.yml for company details, license #, inspector credentials, and โ€” critically โ€” brands_carried and primary_distributor so recommendations tie to parts the shop actually stocks
  • Reference knowledge-base/terminology/ for correct HVAC component terminology
  • Reference knowledge-base/serial-decoders/ for OEM serial-number age-decode patterns
  • For realtor-pre-purchase audience, pull boilerplate disclaimers from knowledge-base/pre-purchase/ (or use the defaults below)
  • Use the company's documentation tone from config.yml โ†’ voice

Serial number decode reference (condensed; full table in knowledge base):

  • Carrier / Bryant / Payne / ICP: First 2 digits = week, next 2 = year (e.g., 2419E12345 โ†’ week 24 of 2019). Post-2005 format.
  • Trane / American Standard: Year is position 2 or 3 (letter- or digit-coded). Older pre-2010 uses Julian-style.
  • Lennox: Last 4 digits of serial = week+year (e.g., ...2419 โ†’ week 24 of 2019)
  • Rheem / Ruud / Weather King: Positions 1โ€“4 = WWYY (e.g., 2419xxxx โ†’ week 24 of 2019)
  • Goodman / Amana / Daikin NA: First 4 = YYMM (e.g., 1906xxxxxx โ†’ June 2019)
  • York / Coleman / Luxaire / Johnson Controls: 10-char serial; year letter in position 4 or use manufacturer lookup
  • Mitsubishi / Fujitsu / LG / Samsung mini-splits: Brand-specific decoders โ€” consult knowledge-base

When serial is visible in a photo, decode it and state the manufacture year with confidence (high if pattern confirmed, medium if partial, low/VERIFY if unclear).

Process:

  1. Catalog each photo/description โ€” Assign reference numbers (Photo 1, Photo 2, etc.) and identify the primary subject
  2. Identify components โ€” List visible components using correct industry terminology
  3. Assess condition โ€” Standard rating scale:
    • โœ… Good โ€” Operating normally, no visible issues
    • โš ๏ธ Fair โ€” Minor wear or early-stage concerns, monitor at next service
    • ๐Ÿ”ด Poor โ€” Requires attention, repair recommended within 30 days
    • ๐Ÿšจ Critical โ€” Safety hazard or imminent failure, immediate action needed
  4. Extract and decode equipment data โ€” Pull nameplate data: model, serial, manufacture date (via decoder), refrigerant type, electrical ratings, capacity (tons/BTU). Where serial is present, show the decoded manufacture date plus the decode method so it can be audited.
  5. Flag safety and code concerns โ€” Code violations, improper installations, safety hazards (missing disconnect, improper flue termination, damaged wiring, refrigerant oil stains)
  6. Compile recommendations โ€” Prioritize by severity and โ€” when proposing replacement โ€” reference config.brands_carried so the suggested replacement is something the shop can actually install
  7. Apply audience-specific boilerplate:
    • realtor-pre-purchase โ†’ add inspection-scope limitations paragraph, inspector credential line, and signature block
    • warranty-file โ†’ include a photo-evidence cross-reference table keyed to the OEM's required fields
    • homeowner โ†’ plain-language summary box up top
    • property-manager โ†’ multi-unit roll-up table and annualized capex implication
    • internal-qc โ†’ strip customer-friendly language, keep just facts and action items

Output format:

VISUAL INSPECTION REPORT
=========================
Date: [today's date]
Inspector: [tech name] ยท [license/cert # from config]
Customer: [if provided]
Location: [address if provided]
Inspection Type: [routine/pre-purchase/warranty/post-install/damage/compliance]
Audience: [homeowner | realtor-pre-purchase | warranty-file | property-manager | internal-qc]

[HOMEOWNER SUMMARY BOX โ€” only when audience = homeowner]
Plain-language summary in 2โ€“4 sentences: what you have, what's fine, what needs attention, what to do next.

EQUIPMENT IDENTIFIED
--------------------
| Unit | Type | Make/Model | Serial # | Decoded Mfg Date (method, confidence) | Age | Refrigerant | Capacity |

PHOTO-BY-PHOTO FINDINGS
-------------------------
### Photo [#]: [Subject Description]
- Components visible: [list]
- Condition: [Good / Fair / Poor / Critical]
- Observations: [specific findings]
- Action needed: [None / Monitor / Repair / Replace / Urgent]

CONDITION SUMMARY
-----------------
| Area/Component | Rating | Priority | Notes |

SAFETY & CODE CONCERNS
-----------------------
[Bulleted list โ€” omit section if none]

RECOMMENDATIONS
---------------
1. [Critical items โ€” immediate action]
2. [Repair items โ€” within 30 days]
3. [Monitor items โ€” next scheduled service]
4. [Efficiency improvements โ€” optional upgrades]
[For replacement suggestions, cite the specific brand/model the shop carries per config.brands_carried]

EQUIPMENT DATA EXTRACTED
-------------------------
[Consolidated nameplate data useful for warranty lookups, part ordering, and system records]

[AUDIENCE-SPECIFIC APPENDIX]
- realtor-pre-purchase: Scope & limitations, inspector signature, liability disclaimer
- warranty-file: OEM cross-reference table (claim portal, serial, failure mode, photo refs)
- property-manager: Multi-unit roll-up, 3-year capex projection
- internal-qc: QC checklist pass/fail summary

Quality standards:

  • Reference specific photo numbers in every finding
  • Use precise HVAC terminology; mark uncertain IDs [VERIFY IN FIELD]
  • Do not invent damage not supported by the photo or notes
  • If a photo is unclear, note "insufficient visibility" rather than speculating
  • Show the serial-number decode method explicitly so anyone can audit the age call
  • For replacement recommendations, cite only brands in config.brands_carried (fall back to brand-agnostic if no match)

Example Output

Given input: "Photo 1: outdoor condenser unit, Lennox XC21, serial 3819F12345 visible on nameplate, coils are pretty dirty with bent fins on the left side. Photo 2: electrical disconnect box next to the unit, wiring looks okay but there's some green corrosion on the connectors. Photo 3: indoor air handler closet, condensate drain line has algae buildup, drip pan has some standing water. This is for the Hendersons, pre-purchase inspection." โ€” Audience: realtor-pre-purchase.

VISUAL INSPECTION REPORT
=========================
Date: 2026-04-14
Inspector: [Name] ยท NATE #[cert from config]
Customer: Henderson โ€” Pre-Purchase Inspection
Location: [Property address]
Inspection Type: Pre-purchase evaluation
Audience: realtor-pre-purchase

EQUIPMENT IDENTIFIED
--------------------
| Unit | Type | Make/Model | Serial # | Decoded Mfg Date (method, confidence) | Age | Refrigerant |
|------|------|------------|----------|---------------------------------------|-----|-------------|
| Outdoor | Split-system condenser | Lennox XC21 | 3819F12345 | Week 38, 2019 (Lennox last-4 = WWYY, high) | ~6.5 yrs | R-410A (era default) |
| Indoor | Air handler | [Not visible โ€” VERIFY] | โ€” | โ€” | โ€” | โ€” |

PHOTO-BY-PHOTO FINDINGS
-------------------------

### Photo 1: Outdoor Condenser โ€” Lennox XC21
- Components visible: Condenser coil, fan grille, service valves, refrigerant lines, nameplate
- Condition: โš ๏ธ Fair
- Observations: Moderate coil soiling, bent fins across roughly 15โ€“20% of the left-facing panel.
- Action needed: Clean coils, straighten fins with fin comb

### Photo 2: Electrical Disconnect
- Components visible: Fused disconnect, service wiring, connectors
- Condition: โš ๏ธ Fair
- Observations: Wiring intact, no arcing damage. Green oxidation (verdigris) on copper connectors.
- Action needed: Clean and treat with anti-oxidant compound

### Photo 3: Indoor Air Handler โ€” Condensate System
- Components visible: Condensate drain line, drip pan, drain trap
- Condition: ๐Ÿ”ด Poor
- Observations: Algae growth in drain line; standing water in drip pan suggests restricted flow.
- Action needed: Flush drain line, algaecide treatment, verify slope

RECOMMENDATIONS
---------------
1. Clear condensate drain blockage (high priority) โ€” prevent overflow and water damage
2. Clean condenser coils and straighten fins (medium) โ€” restore airflow efficiency
3. Clean and protect electrical connections (medium) โ€” prevent further oxidation
4. Schedule comprehensive system test (recommended) โ€” full diagnostics on a 6.5-yr R-410A unit before close
If future replacement is desired, the shop carries [config.brands_carried]. A matched like-for-like Lennox replacement is available at the shop's primary distributor ([config.primary_distributor]).

EQUIPMENT DATA EXTRACTED
-------------------------
- Outdoor: Lennox XC21 ยท S/N 3819F12345 ยท Mfg ~Sep 2019 ยท R-410A
- Indoor: Not visible in submitted photos

APPENDIX โ€” Pre-Purchase Scope & Limitations
--------------------------------------------
This visual inspection is limited to the equipment and areas photographed or accessible at the time of the walk-through. It does not constitute a mechanical test of refrigerant charge, airflow, or electrical load, which require onsite diagnostic equipment. The age estimate for the outdoor unit is based on a high-confidence decode of the Lennox serial number format (positions 5โ€“8 = week + year). A complete pre-close test is recommended for any system over 5 years old.

Inspector: [Name], [License #], [Date]
[Company] ยท [Phone] ยท [Email]

This skill is kept in sync with KRASA-AI/hvac-ai-skills โ€” updated daily from GitHub.