You run the treatment plant, but its history is in shift logs
How do we see what is really happening at every treatment plant we operate, without driving to each one?
Figures and rules on this page apply to
UK and European Union
Working somewhere else? The shape of the problem usually travels. The deadlines do not.
What this looks like
A permit sample fails at a small works. The SCADA trend shows the dip. The reason is in last week's shift log: a blower was nursed along while a part was on order. The operator knew. The manager two counties away found out from the lab result.
The numbers
Every figure here is someone else’s. Check them.
- 30,354urban wastewater treatment plants in operation in the EU
- 2045the year EU treatment plants must be energy-neutral under the revised directive
- 2,801pollution incidents from water company assets in England in 2024, the most since 2011
- 60%rise in serious pollution incidents in England in 2024, to 75
Why it happens
It is not a people problem.
Each plant is run by a small team that writes down what it sees. Logs, checklists, lab sheets and job reports pile up at every site. Head office gets the alarms and the invoices, not the reasons. When a firm grows by winning contracts and buying other operators, each new site brings its own way of writing things down.
Why your current software has not fixed it
Because it was never built to.
SCADA records what the instruments say. The maintenance system records which jobs were opened and closed. Neither reads the operator's note that explains the numbers. Lab results arrive as PDFs. So the story of each plant lives in documents that only the people on site ever read.
Intelligence, plumbed in
An operator's note saying 'blower 2 running hot, part due Friday' is a record. It can be read, filed against the right plant, and raised before the sample fails.
We build the layer that reads every log and report you already keep, then answers questions about each plant from it. Underneath it: a connector, one agreed meaning per field, and a test set scored on your own records.
How we make AI survive real data-
Connectors -
A semantic layer -
Evals you can check
However hard, whatever it is
Treatment plants are one example. Bring any site your teams run on a client's behalf.
We wrote this about water because the pollution and permit figures are public. The same gap exists wherever small site teams write the only honest record.
-
01 We sit with you
Days where the work happens, not a workshop in a meeting room. We watch the job get done and write down the shortcuts nobody wrote down.
-
02 We read everything
Your data, your rules, your vendors and their documentation, and the published research on your sector. We report what is actually in there.
-
03 We break it to first principles
Not which tool fixes this. What is actually causing it, taken apart until we reach the piece that cannot be divided further.
-
04 Then we build
Weeks, not quarters. By this point we are not guessing what to build, and guessing is the thing that makes projects long.
What we build
Specific enough to argue with.
Four mechanisms, not four features. Each one is a thing that happens on its own, every day, whether or not anyone remembers to run it.
-
Shift logs, lab sheets and job reports from every plant read and joined into one running history per site.
-
Early warnings when notes and results point towards a permit breach, before the sample fails.
-
Repeated faults grouped across plants of the same design, so design teams learn from the plants they run.
-
A plain answer to 'what happened at this works last month' for managers who are not on site.
How you would know it worked
Numbers in your own reporting, not ours.
- Days between the first written sign of trouble and a manager hearing about it.
- Permit failures per plant per year.
- Hours spent building monthly client and regulator reports by hand.
Straight answers
Where a model is involved, it is scored against your own records first. Accuracy per source, not one flattering average.
The questions this raises
-
We already have SCADA and a maintenance system.
Keep both. They record numbers and jobs well. This reads the notes and reports that explain the numbers, which neither system uses.
-
Some of our plants still use paper logs.
Scanned or photographed pages can be read too. Those plants often have the most to tell.
-
We keep buying other operators with their own paperwork.
That is where this helps most. Each firm's reports are read as they are, so nobody has to rewrite them into a new format first.
Where the figures come from
We did not make these up, and you should not take our word for them.
Somebody noticed first, and nothing consumed it
The earliest warning is a note in a box nothing reads.
A driver, a fitter, an engineer or an adviser saw it coming and wrote it down. It went into free text. Free text does not trigger anything, does not add up, and never tells you that four people reported the same thing this week.
- ConstructionBuilding productivity rose. Highways and bridges went backwards
- EducationEnrolment is up this year. The intake pool peaked last year
- ManufacturingProductivity fell in fourteen of the twenty manufacturing groups measured
- StaffingFive million people hired and five million separated in the same month
- Transport & fleetMore than a third of crashed trucks already had out-of-service defects
Different work, same box nobody reads. Whatever your people notice first, it is being written down somewhere already.Tell us what yours write down.
The reason a plant is struggling is usually written down already.
We build the layer that reads every log and report you already keep, then answers questions about each plant from it.
See everything we build-
Software -
Hardware -
Ways of working -
Whole ventures
Is this happening to you? Tell us the size of it.
Twenty minutes. We will tell you honestly whether the numbers justify doing anything about it.