Water Damage Restoration Software: Moisture Logs and Scope Notes Checklist

The software should record, for every job, dated moisture and atmospheric readings at fixed points, labelled photos, equipment placed and removed with times, and scope notes tied to rooms and materials. AI can summarise the daily log and tidy scope notes from voice memos, but the technician confirms every reading and every line.

Software is only a container. A platform with an AI summary button will turn a thin file into a polished summary of thin facts, and an adjuster reading it will see the gaps anyway. Fix the file first, then choose the software that makes it hard to leave a file incomplete.

Follow me on Instagram@sagnikteaches

The job file, in the order a job unfolds

Each group below lists what to capture, why it earns its place and how to check it. Trade guidance backs the main idea. One restoration-industry article describes documentation as a safety net for supporting invoices and reducing liability, and it lists authorisation to repair, authorisation to pay and a statement of satisfaction among the papers to hold. The industry's reference standard, ANSI/IICRC S500, covers psychrometry and drying as well as project documentation, and the full text is paid. Use this checklist as a prompt for your own procedures, not as a substitute for the standard.

Connect on LinkedInSagnik Bhattacharya

Arrival: authority, source and the first photos

• Times: first call, arrival and the loss as reported, and who reported it. A timeline shows response speed. Check that the times are system-stamped, not typed afterwards.
• Authorisations: work authorisation, and any authorisation to pay or direction to pay, signed before work starts. Check that the signature time precedes the first piece of equipment placed.
• Source of water: what it was, whether it has been stopped and who stopped it. The IICRC overview notes that finding and correcting the underlying source is the property owner's responsibility, so record what you saw and what the customer was told.
• Affected and unaffected areas: a sketch or floor plan with each room marked, and wide photos of each.
• Pre-existing conditions: old stains, earlier repairs, anything that was there before. Photograph it so new and old damage can be separated later.
• Hazards noted and who was told. This stays a human judgement, because no assistant has stood in the room.

Category and class, with the reasons

Record the water category and the class you assigned, the time, and why. The same trade article advises documenting the reasons for choosing the category on each loss, and describing what you actually saw and smelled instead of pasting the standard's wording. Category can change: a trade explainer on S500 notes that time and temperature affect it, so log the time of each reassessment.

Subscribe on YouTube@sagnikteaches

Class describes initial conditions only. It rests on how much surface is wet and how much water sits in porous or slow-drying materials, so it needs a reason too. A whole room with wet carpet and pad is a different job from one wet corner, and a low-porosity material such as concrete or stone brings specialty drying. Check that the reason field holds a sentence and not boilerplate.

The moisture map and daily readings

• Monitoring points: numbered on a sketch for each affected room, with the material at each (wall, floor, subfloor, insulation, cabinet base). The same points are read on every visit.
• Per reading: point, material, meter used, scale or reading type, value, date, time and the technician's initials. Readings on different scales are not comparable, so the scale matters as much as the number.
• Dry standard: a reading from the same material in an unaffected area. It is the target each point is measured against.
• Atmospheric readings: temperature, relative humidity and humidity ratio (grains per pound or grams per kilogram, whichever your instrument reports) for the affected area, an unaffected area and outside, on every visit. Add dehumidifier inlet and outlet readings if your procedure uses them.
• Photos of the meter display at each point on day one and on the day you call the point dry.

The check is gaps and jumps: no point missing between visits, no change of units, and no unexplained rise. A reading that goes up needs a sentence beside it.

Equipment: what went where, and when

• Per item: type, asset or serial number, room, date and time placed, date and time removed, and the hours on the unit's meter if it has one.
• Reason for every change: "added a second dehumidifier after readings stalled" is worth more than the placement alone.
• Photos of placement in each room, showing the unit and its surroundings.

Check that every unit on the van's out-sheet appears in the job log, and every unit in the log has a removal time. Equipment that is billed but never logged is the first thing a desk reviewer queries.

Scope notes that survive a desk review

Per room: the materials affected (drywall or plasterboard, baseboard or skirting, flooring type, subfloor, insulation, cabinets), what is removed and what is dried in place, with a quantity, a unit and a reason for each line. Add cut heights, disposal, extraction method, containment, contents handling and any cleaning or antimicrobial treatment with the product named. List exclusions and items for other trades, such as the plumber who must fix the source, and note anything the customer declined and signed for. The check is that every removal line carries a reason that points at a reading or a photo.

Daily notes, then completion

Each visit gets a short note: what the readings showed, what changed, what was decided and who was told. Completion needs final readings at every point against the dry standard, removal times for every unit, final photos and the customer's sign-off. A trade article on running water loss jobs advises painting the picture for each loss, in sight and smell, rather than copying the standard. That sentence is also a good test for any field in your software: could a stranger reconstruct the job from it?

What to ask any restoration software before you pay

Restoration platforms fall broadly into job-management systems that run the whole business and field-documentation apps that capture the file on site. Two examples, using what their own pages say. Albi describes itself as restoration management software. Its base plan lists photo documentation, moisture mapping, equipment scanning and e-signatures, there are three tiers (Base, Pro and Enterprise), and no prices are shown. The page also links to an "Albi AI" feature without saying what it does, so ask for a demonstration. Encircle is a field documentation platform. It says it logs psychrometric, moisture content and equipment readings, works fully offline, and uses AI to turn field data into what it calls an IICRC-aligned mitigation scope, which can be sent to Xactimate as a draft estimate for an estimator to review and edit. It describes its pricing as not seat-based and shows no figures.

Whichever you look at, put these questions to the vendor and ask to see the answer on screen.

The strongest test is a trial on a finished job. Enter three days of readings from a real file, add the photos, and see whether you can produce what the insurer asked for, at the speed a technician would tolerate in a wet basement. Software that looks good in a demonstration and slows the technician down will be skipped on busy days, and then the file is thin whatever the software can do.

A kitchen supply-line leak: AI on the daily moisture log

Here is the main worked example. A failed supply line behind a kitchen sink has wet two walls, the floor by the dishwasher and the subfloor under the cabinet base. Four visits, four monitoring points, and a dry standard taken from the same materials in an unaffected room. All figures are illustrative, and the readings are on whatever scale each meter uses for that material.

Point,Material,Dry standard,Day 1,Day 2,Day 3,Day 4
P1,Drywall behind sink,8,38,24,14,9
P2,Drywall left of fridge,8,22,15,10,8
P3,Hardwood near dishwasher,9,26,21,19,17
P4,Subfloor under cabinet base,11,31,27,28,22

Humidity ratio (grains per pound),Day 1,Day 2,Day 3,Day 4
Affected kitchen,92,71,58,49
Unaffected room,55,54,56,55
Outside,62,58,60,57

The prompt does three things: it limits the assistant to the numbers, it names what to flag, and it forbids the conclusion that belongs to the technician.

Summarise the readings below for the daily progress note.
Report only what the numbers show. Flag any point that rose or stayed flat between visits, any point still above its dry standard on the latest visit, and any missing value.
Do not say that anything is dry, complete or on track. The technician decides that.

[paste both tables]

A first draft, as an assistant might return it, reads: "All four monitoring points are trending down and the kitchen is close to dry. Humidity in the affected area fell from 92 to 49. P1 and P2 have reached their dry standards. P3 and P4 are improving." It reads well. It is also wrong in three ways, and each one is a pattern to watch for.

• The forbidden conclusion crept back. "Close to dry" is not supported: P3 reads 17 against a standard of 9, and P4 reads 22 against 11. Delete it, and tighten the prompt if it keeps returning.
• An anomaly was smoothed over. P4 rose from 27 to 28 between day 2 and day 3, so "all four trending down" is false. The note must report the rise, and the technician adds what they know about why.
• "Reached" was rounded up. P2 is at its standard (8 against 8), but P1 reads 9 against 8, one above. The honest wording is "P1 at 9 against 8; P2 at 8, standard met".

The corrected note the technician signs is shorter and plainer: humidity ratio in the kitchen fell from 92 to 49 grains per pound while the unaffected room stayed at 54 to 56; P2 is at its dry standard and P1 is one above; P3 and P4 remain well above theirs; P4 rose on day 3 and the cause is [technician's explanation]. Everything in it can be traced to a number in the table. That is the test for AI text in a restoration file: every sentence points at a reading, a photo or a signature.

From a voice memo to scope notes, with every guess marked

Technicians talk faster than they type, and the first draft of a scope is often a voice memo recorded in the van. The risk is that an assistant tidies it into confident, round numbers. The fix is a prompt that preserves doubt. Here is a memo of the kind you might get for the same kitchen.

ok kitchen so we cut the drywall about two foot up on the sink wall and the left wall, about eight foot run all in I think, maybe more behind the fridge. pulled the kickers, bottoms of the base cabinets are wet. hardwood is cupped near the dishwasher, probably replace. subfloor under it wet but we can dry it. wet insulation out of the sink wall, about four bags. sprayed antimicrobial on the studs.

Turn this voice memo into scope notes by room: material, action (remove, or dry in place), quantity with unit, and reason.
Use only what is said. Where a quantity or height is approximate or missing, copy it as said and add [confirm]. Do not add any work that was not mentioned.
End with a section called "Not stated" listing what a scope normally needs that the memo does not mention.

A good answer comes back as a list: drywall, sink wall and left wall, removed, cut about two feet up [confirm], run about eight feet in total and possibly more behind the fridge [confirm], reason wet readings at P1 and P2; base cabinet kickers removed, bottoms of base cabinets wet [confirm which units]; hardwood near the dishwasher cupped, probable replacement [confirm]; subfloor beneath it wet, dried in place; insulation in the sink wall removed, about four bags [confirm]; studs treated with antimicrobial [confirm product and amount]. Under "Not stated" it lists containment, disposal route, the antimicrobial's product name and photo references.

The technician's job is then mechanical: replace each [confirm] with a measured number, add the photo references and the product name, and delete nothing silently. Compare a vague line with the finished one. Before: "Wet drywall removed in kitchen, equipment set." After: "Kitchen, sink wall: drywall removed from floor to 24 in, 4 ft run, because P1 read 38 on day 1 against a dry standard of 8; photos K-07 to K-09; insulation behind it removed, 4 bags." The second can be checked against the readings and the photos, and the first cannot.

Two shorter jobs for AI: the gap check and the customer update

The gap check compares your file against your own checklist. Give it the checklist and a plain summary of what the file contains, and make it prove absence rather than assume presence.

Here is our job-file checklist and a summary of what this file contains.
List every checklist item for which the summary shows no evidence. Do not assume anything exists because it is usual.

Checklist: [paste]
File summary: [documents attached, photos per room, reading dates, equipment list, signatures]

On the kitchen file it might return: no dry-standard reading recorded for the subfloor; no photo of the P3 meter display on day 1; no removal time for the second dehumidifier; no signed completion statement. It can only compare the text you give it. It cannot know a signature exists unless your summary says so, which is why the summary is built from the system's own attachment list rather than from memory.

The customer update is the other chore. Ask for a message from the day's confirmed note, with a rule against anything the company cannot promise: "Do not mention insurance, cost, cover or completion dates." The output is short: "Hi [first name], we took readings this morning. The walls by the sink are at or very close to their target. The floor near the dishwasher is still wetter than it should be, so the equipment stays running. We will be back [day] to check again. Call me on [number] if anything changes." Read it against the confirmed note before it goes. "At or very close to" is the assistant rounding again, so write the real position, or cut the sentence.

The reading that was quietly corrected

One failure deserves its own section because it undermines everything above. Imagine a technician who pastes the log into an assistant and asks for a tidy summary. One point reads 61 where it read 16 the day before. The assistant assumes a transposed digit, writes 16 in the summary and moves on. The log still says 61, because the technician had re-opened the joint that morning and the rise was real.

When a reviewer later compares summary and log, they disagree, and the question stops being about drying and becomes about which document is true. The rule that prevents this is simple. AI reads the readings table and never writes to it. Summaries live in a draft field, the readings table is edited only through a logged change, and the technician initials the confirmed note. If a tool lets an assistant edit the readings, switch that off.

Insurer documentation: the limits on what AI can be

The record is the readings, the photos, the equipment log, the signed authorisations and the dated notes. An AI summary is a convenience layer on top of that record, and the limits follow from it.

• Insurers set their own rules. Adjusters and programme managers differ on format, timing and platform, and the requirements can change from one claim to the next. Ask before the job starts what they need and how they want it. The tutorial on insurer paperwork for body shops shows the same discipline in another trade.
• Preapproval is a record too. One trade article advises getting preapproval from the insurer for items such as air filtration. Log who approved what, and when, in the file.
• Photos stay untouched. Do not let an assistant relight, enhance or tidy a photo. Once it has been edited it is no longer a record of what the technician saw. Keep the originals with their system timestamps.
• No coverage talk. An assistant cannot say what a policy will pay and should never be allowed to tell a customer. Keep that out of every drafted message.
• The standard outranks the assistant. The IICRC overview notes that a new consensus body has begun revising the 2021 fifth edition, and an assistant will not know which edition you work to. Your procedures and the standard decide; the assistant only helps you write them down.

Estimating has its own limits, covered in what AI can and cannot do with Xactimate estimates. If a job turns up suspected mould, that is a separate report with its own boundaries, and the mould inspection report tutorial shows where AI helps there.

Treat photos as customer data as well. Crop out faces, mail, documents and valuables before any image goes into an assistant. Use a business plan, since ChatGPT Business and Claude Team do not train on business content by default, and read your own customer agreement. When in doubt, give the assistant your typed notes and the readings table instead of the pictures.

Try it on your last finished file

You do not need new software to find out where you stand. Take your most recent finished job, run the gap check from earlier against the checklist, and count the gaps. If the gaps cluster in one group, such as equipment times or dry standards, fix the template or the form first, because a technician who forgets the same thing on every job is telling you the form is wrong. Then sit through two vendor demonstrations with the question table in your hand and the finished file open beside it. For the front of the job, answering and booking flood enquiries with AI covers the calls that start the file.

Questions restoration owners ask about AI and the job file

Can AI decide the water category or class?

No. Category and class are the technician's assessment on site, and the reasons belong in the technician's own words. An assistant can remind you to record both and can compare your notes with the checklist, but a category chosen by software that never saw, smelled or touched the water is not a defensible record.

Is restoration-specific software necessary?

Not for a very small operation, but it saves effort. A shared spreadsheet, a photo app and a PDF template can hold a complete file if everyone fills them in the same way. Restoration platforms put moisture maps, equipment tracking and report export in one place, which matters once several technicians run several jobs.

Is it safe to put customer photos into an AI assistant?

Treat photos as customer data. Crop out faces, documents and valuables, use a business plan that does not train on your content by default, and check what your customer agreement says. If you are unsure, give the assistant the readings table and your typed notes instead of the photos.

Will an insurer accept an AI-written summary?

Insurers and programme managers set their own documentation rules, and these vary by claim. Treat an AI summary as a cover note. The readings, photos, equipment log and signed authorisations remain the record. Ask the adjuster what they need before the job starts, and keep every original file unedited.

Further reads

• Can AI Draft Flooring Contract Clauses for Subfloor and Moisture? — Moisture readings from the flooring side, before and after drying.
• Construction Handover Pack Checklist: Gather the Paperwork With AI — Assemble the closing paperwork when the rebuild starts.
• Gutter Cleaning Invoices With Photo Proof and AI Job Summaries — A simpler photo-proof routine you can borrow for small jobs.
• Water Heater Installation Checklist: Photo Proof and Handover Pack With AI — A leaking tank is a common cause of loss; see the installer side.
• Solar Installation Quality Checklist: Sort Photo Evidence With AI — Another trade that lives on photo evidence and sign-off.
• Leak Detection Reports: AI Turns Findings and Photos Into Customer Reports — A leak detection report workflow: structured field sheet, numbered photos, two AI passes, wording that avoids liability and a check on every figure.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Encircle and Albi product pages; IICRC S500 overview page; trade articles on water loss documentation (cmmonline.com), S500 terms (cleanfax.com) and water damage classes (alorair.com). Checked November 2026.