Home services and trades
Restoration job files (estimates, photos, moisture logs) as AI data
By SourceX Editorial · Updated
Short answer
Restoration job files are useful AI data when the estimate, photos, moisture readings, drying logs and final scope stay linked to one loss. Together they show how damage was assessed, what was done and whether drying succeeded. Homeowner, insurer and program terms shape what can be licensed, so a rights review comes before any export.
Key takeaways
- A restoration job file is valuable as a linked whole: intake, assessment, readings over time, scope changes and closeout.
- Moisture and drying logs are the outcome layer, because they show whether the work met its dry goal.
- Claim numbers, policy details, adjuster correspondence and homeowner identity need removal or a rights check before any outside use.
- Program agreements with carriers or third-party administrators, and estimating software terms, may limit how files are used.
- Photos need their own preparation pass for faces, belongings, documents, house numbers and embedded location data.
What is in a restoration job file?#
A restoration job file is the full record of one loss, from first notice through mitigation, reconstruction and closeout. It is usually spread across a job management system, an estimating platform such as Xactimate, a field documentation app for photos and readings, and email with the carrier or adjuster.
The table lays out a typical file structure, what each part shows, and the rights or privacy note that comes with it. Water, fire and mold jobs vary in detail, but the layers are similar.
| Component | Typical records | What it shows | Rights or privacy note |
|---|---|---|---|
| Intake and authorization | Loss notice, cause of loss, signed work authorization, contact details | How the job started and what was reported | Homeowner identity and signatures; remove |
| Initial assessment | Walkthrough notes, affected areas, category and class of water | The technician's first read of the damage | Company-authored; review notes for personal remarks |
| Photos and video | Before, during and after images, meter screenshots | Visual evidence of damage and progress | Faces, belongings, documents and house numbers need review |
| Moisture and drying logs | Readings by point and visit, atmospheric readings, equipment placement | Whether drying worked and how it progressed | Low personal content; job keys need protection |
| Estimate and scope | Line items, quantities, supplements, final invoice | What was done and how scope changed | Vendor price data and carrier pricing terms may restrict use |
| Carrier communication | Adjuster emails, approvals, denials, supplement requests | How scope was negotiated and approved | Claim numbers, policy details and correspondence; rights check needed |
| Closeout | Certificate of completion, final readings, customer sign-off | That the job met its drying goal | Signatures and personal details; remove |
Why are linked restoration files useful to AI developers?#
Linked restoration files are useful to AI developers because they pair visual evidence with measured readings and a documented decision about scope. Few record types connect what a technician saw, what instruments measured, what work was approved and whether the result met a measurable goal.
That combination supports several kinds of systems: models that read damage photos and suggest affected materials, assistants that draft or check estimates against documented conditions, and planning tools that learn how drying progresses in different materials and conditions. The moisture log is the outcome layer; it shows whether the work succeeded, not only what was planned.
Volume of photos alone matters less. Large folders of unlabeled images with no link to readings or scope are hard to describe and hard to use. A smaller set of complete files, where every image belongs to a job and a room, is easier to prepare, review and license.
The moisture log is where many files break#
The moisture log is where many restoration files break, because readings are recorded inconsistently, kept in a separate app, or stop before the dry goal is documented. A log that cannot be matched to its job, or that skips visits, loses most of its value as a record of outcome.
Many restorers work to recognized industry standards for water damage restoration, and a complete log reflects that discipline. Check your logs for these elements:
- Reading points tied to a moisture map or room sketch, with the same point labels on every visit.
- Material type for each point, such as drywall, subfloor or framing.
- The dry standard or goal for each material, and how it was set.
- Readings on every visit, with date, time and the meter used.
- Atmospheric readings inside the affected area and outside it.
- Equipment placed, moved and removed, by room and visit.
- Final readings showing the goal was met before equipment came off.
- The job number recorded in the documentation app, so the log rejoins the estimate and photos.
Rights: homeowners, insurers and program terms#
Rights in restoration files are more layered than in most trades records because three parties shape them: the property owner, the insurance carrier, and any program or third-party administrator the job came through. Your company usually authored the notes, readings and photos, but authorship does not settle every question.
Homeowner questions turn on your privacy notice, your work authorization wording and the state privacy laws that may apply. Carrier claim numbers, policy details, adjuster names and correspondence are typically removed, and program agreements with carriers or administrators may contain confidentiality or data-use clauses that need review. Unit prices supplied by an estimating platform may be covered by that vendor's license terms, so check the software agreement before including them. This is general information, not legal advice.
Some files are better left out entirely: jobs tied to litigation or disputed claims, jobs where notes describe occupants' health, and losses at sites with security sensitivities. Counsel assesses these questions deal by deal.
Preparing photos before anyone outside sees them#
Photos need their own preparation pass because the camera captures the whole room, not just the damage. Faces, family photos on walls, mail, medicine bottles, documents on desks, license plates and house numbers all show up in restoration images.
Automated tools help but do not finish the job. Presidio, an open-source tool for identifying and redacting personal information in text and images, warns in its own documentation that automated detection cannot guarantee it finds all sensitive information and that additional protections should be used. Human review stays part of photo preparation, and embedded metadata such as GPS coordinates should be stripped from every file.
Illustrative: a water and fire restoration company assembles complete files#
Illustrative: a fictional water and fire restoration company does most of its work on insurance claims, writes estimates in Xactimate, records readings in a field documentation app and tracks jobs in a separate job management system. The owner wants to know whether its water mitigation history is worth a closer look.
A sample of closed water jobs shows a split. Jobs run by the senior crew have moisture maps with consistent point labels, readings on every visit and final readings at the dry goal. Jobs from a busy storm season have photos and estimates but readings on only some visits, and several logs were never exported from the documentation app.
The company scopes only complete files, excludes every job tied to a disputed claim, and asks counsel to review its program agreements with the carriers it works with most. It also changes the field process so the job number is entered in the documentation app at the first visit, which makes future logs easy to rejoin.
How SourceX approaches restoration files#
SourceX approaches restoration files as linked job records whose value depends on completeness: assessment, readings over time, scope and closeout. A fit check uses only descriptions of systems, years of files and record types, so no photos or estimates are shared at that stage.
If a company proceeds, the Rights step of the SourceX five-step transaction covers homeowner notices, carrier and program terms and estimating software terms, and Preparation removes personal and claim identifiers from text and images. The SourceX Evidence Packet records what was included, what was excluded and the owner's release authorization.
Frequently asked questions
Are Xactimate estimates mine to license?
The estimate you wrote is your record, but it contains elements that may not be yours alone: price data supplied by the estimating platform, carrier pricing arrangements and claim details. Review the software license and any carrier program terms, and consider removing or transforming vendor unit prices so scope and quantities can be used without them.
Do fire and mold jobs work the same way as water jobs?
The structure is similar, but the outcome measures differ. Water jobs have readings and dry goals; fire jobs center on cleaning, odor control and contents, and mold jobs on containment, removal and post-remediation verification. Each type needs its own completeness check, and mold files deserve extra review because notes can mention occupant health.
What about contents inventories and pack-out records?
Contents inventories list a household's belongings, often with photos, which makes them highly personal. They are usually excluded or reduced to broad item categories without images or descriptions that identify the family. Structural records, such as rooms affected and work performed, can be kept separate from contents.
Do we need the homeowner's permission to use job photos?
It depends on your work authorization, privacy notice, the state laws that may apply and how the photos are prepared. Images stripped of people, belongings and location details raise different questions from raw images. Counsel should assess this before any photos are included in a package.
Should reconstruction records be kept with mitigation files?
Yes, where your company did both. Reconstruction scope, change orders and final walkthroughs complete the story of a loss and show how mitigation decisions affected the rebuild. Link them by job number, and note when a different contractor handled reconstruction so the file is not mistaken for complete.
Sources
- Presidio is an open-source, MIT-licensed SDK for PII identification and anonymization in text and images, including a module that redacts PII in images. Source
- Presidio's documentation warns that because it uses automated detection mechanisms, there is no guarantee it will find all sensitive information, so additional systems and protections should be employed. Source
Related resources
See if your company qualifies
A short company assessment. No data uploads are needed.