AI photo review for Boom tasks

Overview

Boom is unusually good at making sure a turnover photo gets taken, and like every operations platform it stops there. A catalog item defines the job, a checklist breaks it into subtasks, and any subtask with the Pic option ticked cannot be completed until the person on site uploads a picture. So the photos exist, in volume, attached to the right listing and the right day. What Boom then records is that the checklist was completed and an approver signed the task off. RapidEye reads the pictures underneath all of that, compares each room to how it looked last time, and returns one finding per problem: damage that is new, cleaning that was skipped, something that has gone missing.

How it works with Boom

Boom's operations module is built out of catalog items, and a catalog item carries a division, a topic like Cleaning, a category like Full Cleaning, an approver, a price and a checklist. Automation rules then spawn those items off events, most of them off a checkout. Because every one of those pieces is configurable by you, the useful way to think about the integration is that your catalog is the standard and RapidEye checks the photos against it.

• The checkout cleaning task. The guest's stay sits between the pictures from the last turnover and today's. A burn on the worktop, a chipped basin, a cracked blind slat, a bar stool that has gone, all come back as a finding with both frames and the reservation dates between them. That is the version of a damage conversation the guest cannot argue with.
• The check-out inspection task. Boom raises it automatically or an area manager triggers it from the listing page, and it is meant to be the second pair of eyes on the clean. RapidEye reads it as exactly that, and adds the thing a human inspector never reports: the rooms their own photos did not cover.
• Subtasks duplicated per bedroom and bathroom. Boom can replicate a subtask by the number of bedrooms or bathrooms a property has, which means a four-bedroom home produces four bedroom photo sets rather than one. RapidEye keeps a baseline per room, so the fourth bedroom is compared to the fourth bedroom and not to an average of the house.
• Subtasks that only apply if the property has the amenity. A hot tub, a grill, a pool gate, a boat lift. Boom only shows those subtasks on the properties that have the amenity, and RapidEye holds those specific items to their own standard: the cover latched, the drain clear, the gate shut.
• The maintenance service call. Where a task was raised off a guest message with the guest's own media attached, RapidEye reads that alongside the turnover set, so the state of the repair is on file with a date instead of living in a chat thread.
• Standards that are yours, not ours. The long description on a catalog item, the wording of a subtask, the inventory a listing is supposed to hold. Whatever you wrote down as the standard is what the photo gets checked against.

A finding names the listing and the room, the kind of problem, how serious it looks, the before and after photos, and a suggested next step. The RapidEye dashboard holds every finding, and a Slack channel or inbox can receive them as well. What happens next in Boom is yours: a new service call, a note against the reservation, a charge. Photos are the starting point rather than the limit. Teams who film a walkthrough instead of shooting stills, teams with a 3D scan of each unit, and teams who want findings scored against an inspection form of their own get that set up with us directly.

How to connect Boom to RapidEye

1. The automation rule behind your turnovers. In the Operations module, open Automations and confirm the rule that creates a cleaning or inspection task on checkout. Boom ships Checkout Cleaning and Check-out Inspection as pre-made automations you can activate. That task carries the listing and the date, and it is what a finding gets tied back to.
2. Photos on the checklist. In the catalog item behind that task, the subtasks with Pic ticked are the ones a cleaner cannot complete without uploading a picture. Those pictures are what RapidEye reads. If a room has no subtask asking for a photo today, either add one or use a RapidEye capture link instead: any phone, no app, a fixed set of angles per room.
3. Add RapidEye as a user. In Boom, open the Main Menu in the top left, then Configuration, then Settings, then the Users tab, and click Add User. Add Rapid Eye with the email support@rapideyeinspections.com on a role that can reach the Operations module. Boom's documentation notes that only a user with the users-management ability can do this.

That is the setup. Cleaners keep My Tasks, the Operations calendar keeps its rules and its approvers, and findings start arriving as tasks close.

What a Boom checklist can prove, option by option

Checked against Boom's help centre and product pages, September 2026.

The interesting part of Boom's operations module is not the task, it is the catalog item behind it. According to Boom's help centre documentation (boomnow.com), the catalog is where you define your tasks, and each item carries a division, a main topic such as Cleaning, a sub-topic such as Full Cleaning, a long description, an approver whose sign-off is required on completion, whether the cost lands on the owner or the company, a price, and a checklist. That checklist is a set of subtasks, and each subtask has its own options. Read them side by side and you can see precisely where the evidence in a Boom account stops.

Subtask options as documented by Boom. The third column is what each one settles by itself, before anyone looks at a picture.

How a turnover task gets created and closed

Boom's automation rules generate tasks from triggers, and the documentation is explicit that a checkout creates a new cleaning or inspection task. A rule can be an individual property rule or a master rule that covers the whole portfolio including properties added later, it can exclude cases such as while a guest is staying, it can fire after a set number of days without a cleaning task, and once saved it generates tasks fourteen days ahead on a rolling basis. Boom ships two of these already written.

The first three stages are Boom's, as documented. The fourth is what the integration adds alongside them.

Does the Pic requirement mean somebody looked at the photo?

No, and Boom does not claim it does. The option is a completion gate: it stops a subtask being ticked until an image is uploaded, which is a genuinely strong design and the reason a Boom account is a good place to run this. What follows the upload is an approver, and an approver is a person approving a task, usually from a phone, often after the crew has left the property, across every unit that turned over that day. At twenty properties somebody might skim. At two hundred nobody does, and the pile that nobody opens is where the burn mark, the missing towel set and the light that never came on stay hidden until a guest writes a review. RapidEye does the opening. Every picture, every task, compared with how that exact room looked the last time it was photographed.

What do contractors see, and does any of this change for them?

Nothing changes for them. Boom's documentation says a user on the contractor role reaches only the My Tasks section and nothing else on login, that My Tasks shows the tasks assigned to that user tailored for the mobile app, and that a manager can drag tasks between assignees and time slots in the Contractors view on the Operations calendar. All of that is untouched. The crew opens the same section, works the same checklist, uploads the same pictures. The only difference is on your side, where the pictures now come back read.

Does this help with a damage claim?

It helps with the step before the claim. Boom's own description of its damage protection has the chain starting with damage being reported and a photo attached by housekeeping, after which the booking data is compiled and the recovery runs. Every part of that is sound except the first link, which depends on a tired person at the end of a turnover noticing a mark and deciding it is worth raising. RapidEye is what removes that dependency: the change is found by comparing the room to itself, dated, and matched to the stay it happened in, so the claim has something to start from whether or not anyone flagged it.

Boom integration FAQ