Today
A plating bath changes with every amp-hour: metal content, conducting salt, brightener and pH drift, and foreign metals build up. The lab's analyses, the dosing log at the line and the coating thickness measured per lot sit in separate spreadsheets and on paper, and topping up is often done by experience. So a bath that has drifted out of its operating window usually shows up only as scrap: coating too thin, stains, poor adhesion or a failed salt spray test.
An example
At Rabenhammer Präzisionsteile the zinc-nickel line is running brackets H-2207 for an automotive customer, specified at 8 to 12 µm with 12 to 16 % nickel. The agent links Tuesday's bath analysis with 18,400 amp-hours since the last top-up and the X-ray readings of the last 30 lots: the nickel content of the coating falls from lot to lot and will reach the 12 % limit in about 16 hours. It proposes adding 14 liters of nickel concentrate and 6 liters of complexing agent and shows the calculation from analysis, bath volume and consumption per amp-hour. The plating shift lead releases the top-up, and the next lot is back at 14 % nickel.
Your data, wherever it lives
How a plating bath is doing is spread out today: analyses in the lab and in spreadsheets, the dosing log on paper at the line, amp-hours and temperatures in the line control, coating thickness in the X-ray instrument, lots and jobs in the ERP. demi reads, understands and links them per bath and per lot, with your rights. The AI sees a drift before it makes scrap and works out the top-up, and your shift lead decides.
Where the data lives today
- Databases
- Machines & sensors
- Scans & paper
- Office files
- QM system
- MES
- ERP
How demi solves it
Three layers in this order: connect the data first, then the rights, then the AI.
- 1 · ConnectData Layer
Data sources in this case: bath analyses from the lab, amp-hour counters and bath temperatures from the line control, the dosing log at the line, coating thickness and alloy content from X-ray fluorescence, and lots and jobs in the ERP and MES. demi replaces nothing and links every analysis with its bath, the amp-hours since the last top-up and the lots that have run through the bath since. For each bath you see where it sits in its operating window and where it is heading.
- 2 · PermitPermission Layer
Bath values, analyses and dosing proposals are visible to the plating shop, lab and quality; the bath suppliers' process specifications only to process engineering. Nothing is added until the shift lead or the head of the lab releases the proposal; every release is recorded with quantity and time.
- 3 · ActIntelligence Layer
The agent works out from analysis, bath volume and consumption per amp-hour how much metal, conducting salt and additive is missing and proposes the top-up. It spots trends across analyses and thickness readings and warns before a value reaches the limit of the operating window. It computes the quantities exactly from the data and the bath supplier's specifications; the language model only explains the proposal.
How it runs
- 1
The lab analyses the bath; the line control supplies amp-hours and temperature.
- 2
The agent links the analysis with amp-hours, dosing log and the thickness readings of the lots since the last top-up.
- 3
It calculates the missing quantity of each component and shows where the bath is heading at the same throughput.
- 4
If a value is heading for the limit of the operating window, it warns the plating shop and the lab before scrap occurs.
- 5
The shift lead checks the proposal, releases the top-up and it goes into the dosing log.
What changes
Baths stay in their operating window instead of being topped up by feel
Less scrap and rework from wrong coatings
Analyses, top-ups and coating values per lot in one record
What does it bring you?
Check it with your own numbers
Every starting value is an assumption you can change. The calculation sits right below, step by step.
Your numbers
Result
+72 hfreed per year
Assumptions and calculation
demi links analysis, amp-hours and coating values and works out the top-up; you release it.
20 min → 14 min per case
How we calculate
60 bath analyses × 20 min × 30% × 1.0 (manual) × 12 months = 4,320 min = 72 h
Value per year, before demi costs
Freed per year+72 h
Assumption: labor cost of $48.62 per hour in US manufacturing in June 2026 (BLS), plus workplace, IT and overhead.
= Value per year, before demi costs
+$4,680
gained per year
An estimate, not a promise. The starting values are assumptions for an industrial company with 150 to 500 people.
Combine with more cases and get the packageQuestions about this use case
How does demi know the target values of my bath?
From your bath supplier's data sheets and your own work instructions: operating window, consumption per amp-hour and dosing rules per component. Where your experience values differ, yours apply, and the agent shows which source it uses.
Can demi control the dosing pumps directly?
The agent proposes quantities; the shift lead decides whether to add them. Where your line already doses automatically, demi checks the dosed quantities against consumption and reports when a pump delivers too much or too little.
Does demi also detect contamination in the bath?
Yes. The agent tracks foreign metals, carbonate or breakdown products from the analyses just like the main components. If they rise, it shows which lots could have been affected since and proposes a Hull cell test, a cleaning or a partial renewal of the bath.
Does this help in a customer audit of the plating shop?
Yes. For every lot, bath analysis, top-ups, temperature and thickness readings sit in one record. If a customer asks about a lot from last year, you find the data in seconds instead of in three binders.
