How it works
From the contract to proof that all of it was done
Doneness starts with the job: the contract, the site and the work it covers. It works out what evidence proves each task, manages each visit to completion, and flags anything not done to contract to the designated manager.
Understand the job
We take in the contract and the scope of work, and map each site: boundaries, areas, entrances and anything that needs special care. Doneness then knows all the work each visit must cover, and what done looks like.
Decide what proves each task
For every task we agree the evidence that proves it was done: photos from set positions, a video of a long edge, or a mix.
Guide the crew on site
The crew checks in at the site and is asked for each piece of evidence, one task at a time, in the app or through the system they already use.
Manage the visit to completion
Each photo and video is checked as it arrives, and the visit is assessed as a whole against the contract: whether every area was covered, whether photos are repeated or reused, whether last visit's flags have been dealt with, and whether the right process was followed. Misses come back to the crew with what to do, so they're fixed before the crew leaves.
Flag problems to the right person
Anything not done to contract, or that can't be judged from the evidence, goes to the designated manager with the evidence and the reason. Your team makes the call.
Keep the record and report it
Every visit becomes a proof-of-service record, kept as long as your region requires. Approved records appear on each client's dashboard.
What gets checked
The whole visit is assessed, as well as each photo
Checking that a photo shows a cleared walkway is the easy part, and on its own it can't prove the job. Doneness looks at everything from the visit together: every photo and video, when and where each was taken, the site map, the contract and the history of the site.
Does it show the task done to standard?
Every photo and video is checked against what done means for that task.
Was every area evidenced?
Evidence is placed on the site map, so an entrance, bed or aisle with no proof shows up as a gap.
Are the photos distinct and from this visit?
Near-identical shots, and photos reused from earlier visits, are spotted and set aside.
Has last visit's problem been fixed?
Anything flagged before is carried forward and checked again until it's done.
Was the work done the right way?
Steps are checked in order, such as pre-treating before a storm, or cutting back and cleaning a pothole before it's filled.
Do the facts fit the work?
Timestamps, GPS and time on site are compared with the work claimed.
The result is one assessment of the visit, with the reason for every flag, sent to the crew while they're on site and to the manager when it needs a decision.
A common question
Why not use a general-purpose AI?
A general model can describe a photo. It doesn't know your contract, your site or what done means for this task.
It knows the contract
Each check is made against what was agreed for that site and service.
It knows the site
Evidence is placed against the site map, so missing areas are found.
It's trained for the work
Small models fine-tuned for specific jobs, like snow clearance and grounds care.
It explains itself
Every result comes with the reason, and unclear cases go to a person.
Where it fits
Alongside the systems you already run
Work orders can come from your CMMS or field-service system, and results go back to it. Crews can use the Doneness app, or keep using the tools they have.
Talk to us about your sites
Tell us how many sites you manage and which services you run. We'll show you Doneness on your own work.