Today
If your product is not in the consultant's specification, you will later negotiate on price alone. Consultants like to take over finished texts, but yours exist only for the main sizes, as old text files with outdated references to standards. Writing new texts by hand for every variant, every option and every change in a standard is more than product management will ever manage.
An example
Nortavo Gebäudetechnik offers its ILZ air handling unit in four sizes from 2,000 to 8,000 m³/h, with plate or rotary heat exchanger, heating and cooling coils, for indoor or outdoor installation, 96 variants in all. Until now, specification texts existed for only two sizes, as text files from 2019 with filter classes under the withdrawn EN 779. The agent writes a neutral and a product-specific text for every variant, for example for the ILZ 40: 4,000 m³/h, heat recovery 78%, supply air filter ISO ePM1 55%, hygienic design to VDI 6022, and generates the GAEB files. In the neutral text it replaces Nortavo's own name for the rotary exchanger with performance data. The product manager checks twelve samples and approves, and an engineering consultancy planning a school refurbishment takes the items for the ILZ 40 straight into its tender documents.
Your data, wherever it lives
The facts for your specification texts are scattered today: variants and performance data in the PLM and ERP, dimensions in CAD, data sheets and test certificates as PDFs, old texts as Word files in folders and on the website. demi reads, understands and links them, with your rights. The AI writes texts and GAEB files for every variant in minutes instead of days, and your product management approves what consultants receive.
Where the data lives today
- PLM
- ERP
- CAD & drawings
- PDFs
- Office files
- Folders & drives
- Web & portals
How demi solves it
Three layers in this order: connect the data first, then the rights, then the AI.
- 1 · ConnectData Layer
The case needs data from the PLM or PIM, ERP and CAD, your data sheets and test certificates as PDFs, and your existing specification texts. demi replaces nothing and links performance data, options, dimensions, standards and evidence for every variant. demi knows which values carry a text and which certificate stands behind each one.
- 2 · PermitPermission Layer
Only released product data goes into the texts; prices, costs, suppliers and development work stay internal. Product management approves every text before it goes on the website, into portals or to a consultant, and which version was published when is logged.
- 3 · ActIntelligence Layer
For every variant the agent writes a product-neutral text that describes performance and properties, and a product-specific one with make and type, each with short and long text, unit and references to standards. In the neutral text it flags wording that points to a particular make, because public-sector clients must tender product-neutrally. From this it generates GAEB files and the design data consultants expect.
How it runs
- 1
demi reads your product data, data sheets, test certificates and existing specification texts.
- 2
The agent writes a product-neutral and a product-specific text for every variant and option.
- 3
It checks values and references to standards against product data and certificates and flags whatever is not backed.
- 4
Product management checks samples and approves.
- 5
The texts are published as GAEB files and in the templates of the specification portals; when a value changes, the agent proposes the new version.
What changes
Consultants find a finished text for every variant
Neutral texts that public-sector clients can use as well
Product changes reach texts and portals without delay
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
+96 hfreed per year
Assumptions and calculation
demi writes the texts from your product data and generates the files; product management approves.
60 min → 36 min per case
How we calculate
20 specification texts × 60 min × 40% × 1.0 (manual) × 12 months = 5,760 min = 96 h
Value per year, before demi costs
Freed per year+96 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
+$6,240
gained per year
96 h ÷ 1,760 h per role = 0.1 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
What is the difference between a product-neutral and a product-specific text?
The product-neutral text describes performance and properties without naming make and type, so a public-sector client can use it too. The product-specific text names your product and suits private clients. The agent writes both from the same data, so they never drift apart.
In which formats does demi deliver the specification texts?
As GAEB files, the German exchange format, for import into tendering software, as text and PDF versions for your website, and in the templates of the specification portals you supply. Design data such as CAD and BIM models and product data to VDI 3805 come with them, where they are usual for your products.
How does demi know the values in the text are right?
Every value in the text comes from your product data and points to its source; the agent checks references to standards against the data sheet and test certificate. If a value is found nowhere, it does not write it but shows you the gap. When a value changes, it proposes the new version of every text affected.
Does this work for other countries and languages too?
Yes. The agent writes the texts in the languages and with the references to standards you define, for example for Austria or Switzerland. You store each country's usual formats once; product management decides which markets are served.
