Daily drying logs that survive an adjuster review.

A drying log is the evidence behind a mitigation invoice: daily atmospheric and moisture readings, the equipment that was running, and photos that tie the two together. Gaps have ordinary causes: a technician skips a room that looked dry, a Sunday visit is cut short, an air mover is moved and the log is not updated. Found the next morning, a gap can still be closed on the next visit. Found at review, it is a line item somebody has to defend. EPA's mold guide, checked on 13 October 2026, advises that water-damaged areas and items be dried within 24 to 48 hours to prevent mold growth; a log showing that drying has stalled is worth finding the next morning, not at review.

This article sets out what the log needs to show, the gaps a morning check looks for, and how an AI assistant can run that check and compare the estimate with the logs, without ever deciding category, class or when the job is dry.

Accuserve, a property-claims and repair company, published two pieces by a restoration trainer (checked on 13 October 2026) that set the bar clearly. One quotes the S500 standard's 2021 edition on recording and monitoring moisture measurements daily until drying goals are met, and criticizes skipped monitoring, weekends especially. The other lists what a drying record should contain: who took the readings and with what instrument, drying goals, temperature and humidity in affected and unaffected areas and at dehumidifier inlets and outlets, moisture maps labeled by room, wall, material and day, and photos of initial readings and the equipment set-up.

Two terms matter for the morning check. The National Weather Service's Louisville office, checked on 14 October 2026, explains that relative humidity measures water vapor relative to the air's temperature, so the same amount of moisture gives a lower relative humidity in warmer air. Grains per pound (GPP) is a weight measure of the water actually in the air, so warming a room can lower relative humidity while GPP stays the same, which looks like progress and is not. A log that records temperature, relative humidity and GPP together lets a reviewer see whether moisture is really leaving.

Moisture readings are only comparable if they come from the same marked points each day. A reading point that appears on day one and vanishes on day three is a gap, even if the room was visited.

The gaps below are the ones a morning check is designed to find. None of them needs judgment to spot; all of them need a technician to fix.

In Grok Bot, a routine tells one Bot when to run a workflow on a schedule. Schedules must be at least five minutes apart, and the app keeps the twenty most recent run records for each routine, which is where the owner checks that the morning run happened. A documentation Bot with one routine, running early every day including weekends, is enough for this job.

The Bot needs to read yesterday's logs. Restoration drying software has no documented connector, so the input is either a custom integration Aevrion builds against the vendor's API or export, or supervised browser use of the vendor's site where the vendor permits it, subject to testing and to human sign-in steps. A nightly export to a shared folder is often the simplest start. The logs stay in the drying software; the Bot's memory is not the record, and the documentation says as much.

1. 01

   Owning Bot and schedule

   Decides

   Which Bot runs it, at what time, in which time zone?

   Output

   The documentation Bot, early each morning, weekends included

2. 02

   Input source

   Decides

   Where do yesterday's readings come from?

   Output

   The export or integration from the drying software, never memory

3. 03

   Expected result

   Decides

   What should exist after the run?

   Output

   One gap list per open job, each gap linked to its record

4. 04

   Approval boundary

   Decides

   What may the routine do beyond reading and listing?

   Output

   Nothing: it prepares the list and sends no message

5. 05

   Missing source

   Decides

   What if the export is absent or stale?

   Output

   Report that the check did not run; never reuse yesterday's data

Based on the Grok Bot documentation at docs.x.ai as checked on 13 October 2026. Its statements on Bot rules, Bot-to-Bot messaging, routine triggers, connected apps, approval expiry and payments were rechecked against docs.x.ai and Cursor's help pages on 26 October 2026. The trade examples are Aevrion's illustrations, not documented Grok Bot use cases. Aevrion Ops is an independent consultancy. It is not affiliated with, endorsed by or a partner of xAI, SpaceXSI or Cursor. Grok Bot is named here only to describe it.

For consequential work, Grok Bot's documentation advises asking a Bot to separate facts found in source systems from assumptions, completed actions, actions awaiting approval and unresolved questions, and to keep source links, screenshots and a short action log. Applied here, each line in the gap list says which job, which room, which point, which day, and links to the record. An assumption, such as 'the dehumidifier was probably moved', is labeled as one.

The routine stops at the list. Messages to technicians are drafts the owner sends, and approvals raised by an unattended routine expire after about ten minutes (Cursor's help pages say thirty) without running, so a prepare-only routine is the design that works at six in the morning.

When drying is signed off and the estimate is drafted, the same Bot can run a second check: equipment lines against the equipment log (units, rooms, days on site, pull dates), drying days against the daily readings, and demolition and antimicrobial lines against the photos and notes. Each difference goes on a list for the estimator, with the records on both sides.

The Bot does not change the estimate. A difference can mean a log error, a billing error or a line that is right and needs a note, and only the estimator can tell which. Texas adds a further line for anyone drafting carrier correspondence: Texas Insurance Code §4102.163 (as published on 13 October 2026) bars a contractor from acting as a public adjuster for property it works on, so carrier drafts stay factual and never argue coverage.

Category and class are the technician's determinations. Category can change with time and temperature, as a Cleanfax explainer on the S500 terms notes (checked on 13 October 2026), so a note on day three that suggests contamination goes to the technician as a question, never as a reclassification. The detailed class definitions sit in the S500 standard itself, and the assistant does not interpret them.

• NEVER FILL A READING

  A missing reading is reported, not estimated, interpolated or copied from the day before.

• NEVER DECLARE DRY

  Whether the drying goal is met is the technician's sign-off.

• NEVER MOVE EQUIPMENT

  Adding, moving or pulling equipment is a site decision.

• NEVER EDIT THE ESTIMATE

  Every line and supplement is the estimator's.

• NEVER CONTACT THE CARRIER

  Carrier messages are drafts a person sends, or are not sent.

Run the check on past jobs first: give the Bot a week of exported logs from finished jobs and compare its gap list with what an experienced technician finds. Fix the procedure where they disagree, then schedule it. The guide to reliable home-service workflows covers routine design in general, including the no-data rule and checking run history.

Aevrion Ops can set this up for a restoration company: the gap definitions, the routine, the estimate check and the integration with the drying software, if the vendor's API or terms allow it, and once it has been tested. Contact us to scope it.

1. A drying log needs daily temperature, relative humidity and GPP readings and moisture readings at the same marked points, with the equipment that was running.
2. A morning routine finds rooms not revisited, missing atmospherics, lost reading points, equipment mismatches and unproven readings while the job is open.
3. The routine only prepares a gap list; unattended approvals expire, and messages to technicians or carriers are drafts a person sends.
4. Before review, the estimate is compared with the equipment log and photos, and the estimator decides every difference.
5. Category, class, the drying goal and equipment changes are human decisions; the assistant never fills in a missing reading.