Today
Refineries, chemical parks and their EPC contractors send an API 610 pump data sheet with every enquiry, plus their own specification that tightens the standard. The data sheet has hundreds of fields: operating conditions, liquid, curve, NPSH, material class, seal plan, bearings, driver, test scope and painting. An application engineer gathers performance curves, material tables and old data sheets for it and easily spends two days per quote. Whatever slips through in the rush, such as a rated point too far from best efficiency or a required seal plan, surfaces only in technical clarification or after the order.
An example
An EPC contractor sends Hartwig & Brunn Pumpenwerk the data sheet for two process pumps in a refinery: 120 m³/h, 80 m head, hydrocarbon at 260 °C, material class S-6, seal plan 53B and a witnessed performance test. The agent picks the pump size from the test-bench curves, fills 286 of 312 fields and takes bearings and driver from a reference order for the same refinery. It flags three deviations: the rated point sits at 76 % of best-efficiency flow, the NPSH margin is smaller than the specification requires, and there is no reference order yet for plan 53B in this size. The application engineer picks the next smaller size, clarifies the NPSH margin with the EPC contractor, settles plan 53B with the in-house seal team and releases the data sheet the next morning.
Your data, wherever it lives
Today the answers to an API data sheet are scattered: test-bench curves and series data in the PLM, material tables and old data sheets in folders and spreadsheets, reference orders in the ERP, and the enquiry with the operator's specification in a mailbox. demi reads, understands and links them per pump and field, with your access rights. The AI prefills the data sheet and writes the deviation list, work that takes two days of searching today, and the application engineer checks and decides.
Where the data lives today
- E-mails
- PDFs
- Office files
- Folders & drives
- PLM
- 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: ERP, PLM, file share, e-mail and customers' data sheets as PDF or spreadsheet. demi replaces nothing and links measured test-bench curves, impeller diameters, NPSH values, material tables, seal selections and earlier data sheets to the orders in which they were delivered. For every field you find the value that has already been right for this pump and this liquid.
- 2 · PermitPermission Layer
Curves and data sheets are visible to sales and application engineering; prices and costing only to sales. A customer's data sheets and specifications are seen only by the team working on that enquiry. The data sheet goes out only after the responsible application engineer releases it; the release is recorded.
- 3 · ActIntelligence Layer
The agent reads the customer's data sheet and specification, scans included, and fills each field from your data, with the source alongside. Curve values, NPSH and power come from measured curves, never from estimates. The check rules, such as the allowed distance from best efficiency or the required NPSH margin, are stored as fixed rules; from them it writes the deviation list, and the language model only words the reasons.
How it runs
- 1
The enquiry arrives by e-mail with the data sheet, the operator's specification and the liquid data.
- 2
The agent picks pump size and impeller from your performance curves and shows where the rated point sits relative to best efficiency.
- 3
It fills material class, seal plan, bearings, driver and test scope from material tables and reference orders.
- 4
Every deviation from standard or specification goes into the deviation list with its reason; open fields are marked.
- 5
The application engineer checks, decides on each deviation and releases data sheet and list with the quote.
What changes
Complete data sheets, every value with its source
Deviations surface before the quote, not at acceptance
Experience from earlier pump orders flows into every new data sheet
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
+288 hfreed per year
Assumptions and calculation
demi fills the data sheet from curves and past orders and flags deviations; you check and release.
10 h → 7 h per case
How we calculate
8 pump data sheets × 10 h × 30% × 1.0 (manual) × 12 months = 288 h
Value per year, before demi costs
Freed per year+288 h
Assumption: labour cost of €49.50 per hour in German manufacturing in 2025 (Destatis), plus workplace, IT and overheads.
= Value per year, before demi costs
+€18,720
gained per year
288 h ÷ 1,760 h per role = 0.2 full-time roles
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
Which edition of API 610 and which operator standards does demi follow?
The edition the customer names in the enquiry, plus the operator's specification that you upload with it. Your application engineering team stores the check rules per edition. The agent shows where an operator standard is stricter than API and then checks against the stricter rule.
Where do curve, NPSH and efficiency in the data sheet come from?
From your measured test-bench curves and the data sheets of your pump series in the PLM. The agent uses released curves only and shows for each value which curve and impeller diameter it comes from. If a size has no measured curve, the field stays open and is marked.
What goes into the deviation list?
Every point where your offer departs from data sheet, standard or specification: field, requirement, your value, reason and source. Questions for the customer come with it, such as missing liquid data. What goes out as a deviation, an alternative or a clarification is the application engineer's decision.
What happens when the customer sends a new revision of the data sheet?
The agent compares it field by field with the previous revision and shows what changed and which of your answers are affected. Changed operating data are checked against the selected pump again. You only look at the affected fields, not the whole sheet.
