One job, from the first form to the finished file

Say your firm has won the lighting audit for a borough council. Every column in the area recorded: where it is, what type it is, how tall, what lamp is in it, what condition the door and the bracket are in, and anything nearby worth flagging. Four people on the ground, three weeks, and a dataset at the end that the council can load straight into their own system.

None of that is the difficult part. The difficult part is everything around it. Four surveyors in four different streets, each filling in what they believe the form is. Four sets of work that somebody has to turn into one at the end of every week. And a client who will change their mind about the form in week two, by which point four phones are already out there collecting against the old one.

An example rather than a case study: this is the shape of the work, not a named customer’s job.

You build the job once. A layer for the columns, another for anything worth flagging along the way, and the fields the council asked for, with drop-down lists wherever the answer should be one of a known set: column type, lamp type, condition. Then you release it, and that is the moment it stops being your draft and becomes the version everybody works to.

Your four surveyors open Mapit GIS and the job is waiting for them, identical on every phone. Nobody has been emailed a file. Nobody is working from last month’s copy with two extra fields somebody added by hand.

Underneath, releasing fixes the structure and gives it a number, and that number is what lets everybody’s work fit back together later.

Your surveyor walks the street, drops a point at each column, works down the form and moves on. Two hundred columns in a day. The drop-downs mean the condition field says one of four things rather than four surveyors’ four spellings of the same word, which is most of what makes a dataset usable at the end.

Signal comes and goes across a borough, and it makes no difference to any of this: the map and the form are on the phone. It syncs when it can.

This is the bit that usually costs a Friday. Four people have collected into four copies of the job, and somebody has to make one dataset out of them without losing anything or counting anything twice.

Here it has already happened. Each phone sends up what it collected and takes down what everybody else collected, so by Friday afternoon there is one dataset and it is the same one on every device. Nobody merges anything by hand. Nobody is working from a copy that is four days stale.

It is also why your surveyor sees the rest of the crew’s work on their own map. The streets already done are obvious, which is the cheapest way there is of stopping two people auditing the same road.

Only the changes travel, never the whole job, so a sync in week three costs the same as a sync on day one.

It happens. One surveyor records column 4412 on Tuesday and marks the door as damaged. Another passes it on Thursday, does not realise it is already done, and records the lamp type, which the first person had left blank.

Both pieces of work land. They filled in different parts of the same record, so the record ends up with both. Had they disagreed about the same field, one answer stands and the other is kept rather than thrown away, waiting in a list for whoever runs the job to look at. Nobody’s afternoon quietly disappears because somebody else got there first.

This is the promise the whole thing is built around. A sync is allowed to make a decision. It is never allowed to make work vanish.

The council asks whether you can also record which feeder pillar each column is fed from. This is the phone call nobody wants. Four phones are out in the field with the old form on them, two thousand columns are already recorded, and on most jobs the honest answers are “from Monday” or “we will add it in the office afterwards”.

You add the field, look at what it changes, and release again. That is the whole job. The next time each phone syncs it picks up the new field and carries on with it, out in the street, that afternoon. The two thousand columns already recorded keep everything they had, and the new field sits empty until somebody fills it in. Nothing is re-issued, nothing is re-collected, and nobody has to be called in.

What you cannot do is rename or change the meaning of a field everybody is already using, and that is deliberate. Adding is safe. Quietly redefining something that has been collected for a week is how a dataset gets ruined, so the product refuses it and tells you to add a new field instead.

A release carries the change itself rather than a new copy of the job, which is why a phone can pick it up on a bus without noticing.

Whoever runs the job has one place to look, and what they are looking at is the job itself rather than a summary of it. Not four files in an inbox, not a spreadsheet somebody stitched together on Thursday: the columns the crew have recorded, as they stand.

On top of that the portal shows the crew: who is working in which layer, how much each of them has collected, and when each phone last synced. Every figure carries the time it arrived, and a quiet row means a phone has not synced yet rather than somebody slacking. The board is careful never to imply otherwise.

Which is what you actually need when the client rings: how much of the borough is done, what has come back from where, and whether the answer is yes.

The audit finishes. You close the job, which stops it taking any more work, and download it as a single file. It is a GeoPackage: an open format that QGIS, ArcGIS and most other serious tools open directly, with the layers, the fields and the drop-down lists intact. One file, not four, and not something only our software can read.

You hand that over, keep your own copy, and delete the job from the cloud when you are ready. Nothing vanishes from anybody’s phone behind their back: their copy simply becomes theirs.

The council’s data is the council’s before it is ours. A job is encrypted whenever it is not actually being worked on, under a key belonging to your organisation and nobody else, and everything stays in the EU.

Nothing is visible to anybody until you share it, which matters when the same firm is auditing lighting for one council and surveying trees for another. And one thing we will not pretend: our own staff with production access can reach organisation data, as they can on any platform that merges your records or serves you a map. What we owe you instead is a record of who looked, so every staff view of an organisation is logged.

We are finalising deployments and starting testing. Mapit GIS Cloud will open to a small number of selected users shortly, and if you have told us you are interested you will be among the first asked.

General availability is planned for January 2027.