Every job closes with its evidence.
Reviews job evidence, inspections, and safety checklists, creates corrective actions, and tracks them through completion. Your office stops opening closed jobs one at a time to find out which ones are missing a photo, a reading or a signature.
- Checks
- Mandatory fields, photos, readings and sign-off
- Works in
- QA Agent ( Internal)
- Runs on
- Dynamic Form Builder and Workorder Management
- Runs
- At close-out, on every completed job
what it is for
The gap turns up months later.
A job is closed in the field and the file behind it is short: a photograph nobody took, a meter reading left blank, a signature the customer was never asked for. Nothing looks wrong at the time. It surfaces when an invoice is queried, when a warranty claim is made, or when an auditor asks for a year of completion records. Then somebody opens jobs one at a time to find out which ones are incomplete.
AI Quality Assistant checks each close-out as it happens. It confirms that what your checklist asks for is actually attached, opens a corrective action where one of your checks has failed, and books the return visit or inspection. It checks that the evidence is there. Whether the repair itself was any good is a judgement, and that stays with a supervisor.
Close-out checks
Complete files, not chased ones
Missing photos and blank readings are picked up while the crew is still on site, rather than weeks later when nobody remembers the visit.
Failures become work, not notes
A failed check opens a corrective action with an owner and a due date, and books the return visit, instead of sitting in a comment on a closed job.
Judgement stays with a supervisor
It reports what is present and what is missing. Deciding whether the workmanship was acceptable is never something it does on its own.
check the close-out
It checks the file before the job closes.
The check runs at the moment of close-out, while the crew is still on site and the fix is still cheap. What it looks for is whatever your own checklist made mandatory.
Fields, photographs, meter readings, serial numbers, certificates and the customer signature. The list comes from the checklist you built for that job type, so the agent is never checking against a standard you did not set.
Where you have given a check a range, a value outside it is a failure and a value inside it is not. The agent is comparing a number with a number, which is the only kind of judgement it makes.
Each check can be set to hold the job open until the item is attached, or to let the job close and raise the gap behind it. Safety-critical items usually hold; a nice-to-have photo usually does not.
Forms filled in a basement or a plant room are checked when the phone syncs, not before. The crew is not asked for an item they have already attached and the job is not flagged twice.
Evidence check
what counts as a failure
You define the failure. It applies it.
The agent holds no opinion about what good looks like. It applies the rules you wrote, in the same way on every job, and sends anything that does not fit them to a person.
An item that was never attached and an item that was attached and fell outside your range are two different outcomes, treated differently, because one is chased from the crew and the other opens corrective work.
The agent can see that five photographs are attached. Whether they show a clean braze or a tidy cable run is not something it rules on, and a supervisor is the one who opens them.
A score below the threshold you set is raised to the service manager with the job attached. The agent sends no reply to the customer and records no view about who was at fault.
Where the same check fails on the same asset or the same job type more than once, those jobs are grouped, so a pattern is visible without anyone running a report to find it.
Approved actions
open the corrective action
A failed check becomes work with an owner.
A finding that sits in a note changes nothing. Each failure becomes a corrective action with an owner, a due date and a visit, and it is a person who signs it off as closed.
Linked to the job, the asset and the customer, with the failed check, the evidence that was attached and the evidence that was not, so the next person sees the case rather than a reference number.
Through your normal scheduling rules, with the qualifications the revisit needs, so a corrective action does not wait for someone to remember to raise a job for it.
The result is logged against the technician or subcontractor who did the work, as your quality process requires. It is a record of what happened on a job, and it is not a rating the agent calculates.
The agent can gather the evidence, chase the missing item and show that the revisit happened. Signing the corrective action off as resolved stays with whoever you have named for that.
Opened this week
how it works
Four steps, then it hands on.
The same four run on every completed job. Step three creates work, and a person is the one who closes it.
-
audits
Verifies that mandatory inspection fields, before/after photos, and client signatures are completed before a job is closed.
-
identifies
Detects failed checklist items, low client satisfaction scores, and photographic defects against quality standards.
-
generates
Automatically creates corrective action tickets and schedules remedial visits or supervisor inspection work orders.
-
records
Logs quality performance scores against cleaner and subcontractor profiles to inform future roster allocations.
where the line sits
What it handles, and what it hands over.
Every closed job lands on one side of this line or the other, and the checklist that decides which one is yours.
| Closed job | What the agent does | Who finishes it |
|---|---|---|
| Complete, nothing missing | Confirms the checklist is satisfied, closes the job and writes the evidence list to the record. | Closes unattended |
| Photo or reading missing | Lists exactly what is short and asks the crew for it while they are still on site. | Technician attaches |
| Checklist item failed | Opens a corrective action against the job and asset, and books the return visit. | Supervisor approves |
| Customer sign-off missing | Holds the job open where you set it to, and passes it to the office with the contact attached. | Coordinator chases |
| Satisfaction score below par | Flags the job to the service manager with the close-out attached. It sends no reply. | Service manager |
| Anything it cannot place | A job closed against a checklist it does not hold is passed on rather than marked complete. | Quality lead |
you set
Before it runs its first job, three things come from you.
- Digital checklist templates with mandatory fields and photo validation criteria.
- Quality failure thresholds that trigger corrective work orders.
- Remediation protocols and supervisor notification rules.
None of it is a data migration. It reads the QA Agent ( Internal) records you already keep.
stays with your team
Enforces strict evidence completeness and initiates corrective workflows automatically. Qualitative assessment of subjective finishes remains backed by supervisor review.
What reaches the supervisor is the job, the checklist, the items attached and the one that failed, so the call on the work itself takes a minute.
rollout
Start with one checklist.
Most teams do not switch every action on at once, and nothing here requires it. The usual order:
Watch first
Point it at one job type with every action switched off. It reports what a week of close-outs was missing, and holds nothing open.
Chase the missing items
Turn on the request back to the crew. Nothing is held, no action is opened, but photos and readings start arriving before people leave site.
Then open the actions
Let it hold the safety-critical items and open corrective actions for one job type, review a fortnight of them, then add the next checklist.
works with
What it runs on, and what it hands to.
read next
What an auditor asks you for.
before you switch it on
What teams ask about AI quality assistant.
Only inside the rules you set. Anything outside them - listed under "you set" above - goes to a named person on your team, with the record attached. There is no state where it silently makes a judgement call you have not already approved.
Days, not months. It reads from the records you already keep in QA Agent ( Internal), so there is no separate data migration or new system for your team to learn before it can start.
Yes, instantly, and every action it took stays on the record so you can see exactly what it did and reverse it. Turning it off does not remove or hide anything it has already written.
No. It takes the routine, repeatable part of the job off their desk so they spend their time on the exceptions and judgement calls that actually need a person - the work listed under "stays with your team" above.
AI quality assistant runs on your QA Agent ( Internal) records, alongside Dynamic Form Builder and Workorder Management, so nothing needs re-keying between systems.
No, and it is not built to. It confirms that the evidence your checklist requires is attached, and it compares a reading with the range you set, because that is arithmetic. Looking at a photograph and deciding whether the repair is acceptable is a supervisor call, and the agent hands the job over for it.
Only for the checks you set to hold, and most teams start with none. Every other check lets the job close and raises the gap behind it. Where a phone was offline, the check runs on sync, so nobody is blocked in a basement by a form that could not upload.
It is a record of what happened on a job, kept because your quality process asks for it. The agent calculates no rating and ranks nobody. Where the same check fails repeatedly it groups those jobs so the pattern is visible, and what is done about it is a management decision.
See it on your own close-outs.
Book a demo
We will run it against your own checklists, a sample of closed jobs and your own rules for what counts as a failure.
Book a demoSee the other nine
AI quality assistant is one of ten agents on one job record, from first enquiry to paid invoice.
All AI agentsSee the platform
The field service system underneath: scheduling, work orders, assets, invoicing and the field app.
AI Field Service Management