Industry
In mechanical engineering, the manual work is not on the shop floor but between sales, design and service.
A special machinery builder with 450 employees: proALPHA as ERP, a PLM alongside the CAD, Salesforce in sales, a service portal for customers. Production runs like clockwork. Meanwhile, a sales engineer reads an 80-page specification, a service colleague searches for the 2009 bill of materials that matches a photo of a nameplate, and technical writing waits for the twelfth language version of an operating manual. This is work that ties up specialists without needing their expertise. AI in mechanical engineering starts exactly there: a workflow reads, searches and translates, the specialists decide.
Workflows
Five workflows that pay off in mechanical engineering
Not a list from a catalogue, but the workflows we find again and again in machine and plant builders. The analysis shows which of them pay off for you, with a column for those that do not.
-
Pre-checking specifications and enquiries
Exchange, CRM, ERP with order history
A plant builder typically receives several dozen enquiries a month, some of them as specifications of 40 to 150 pages. The workflow reads every document, extracts requirements, standards, deadlines and contractual penalties, and places them next to comparable orders from the ERP. It marks deviations from the standard and notes where each one appears. The sales engineer decides on feasibility, not the workflow.
Sales and communicationService -
Answering spare parts enquiries in the service department
Service portal, email, ERP bills of materials, serial numbers
Customers send a photo of the nameplate, an old part number or just a description. A company with several thousand machines in the field often handles hundreds of such enquiries a month. The workflow identifies the machine by its serial number, finds the part in the bill of materials as delivered, checks successor parts and gives the sales office a draft quotation. If the match is not clear, it asks instead of guessing.
Sales and communicationService -
Cleaning up the item master in the ERP
ERP, PLM, purchasing data, standard parts catalogues
After twenty years, an ERP with 100,000 items or more contains the same cap screw in five spellings, bought-in parts without attributes and dead entries from projects nobody builds any more. The workflow detects duplicates by attributes, supplier and drawing number, standardises names according to your rules and puts borderline cases forward for review. Every change is logged and can be reversed. After that, the rules apply to every new item as well.
Putting data in orderService -
Translating and maintaining technical documentation
Content management system, translation memory, PLM, terminology
Operating manuals, spare parts catalogues and data sheets typically go out in ten to twenty languages, and every design change triggers a new round. The workflow translates only the modules that changed, sticks to your terminology and checks every version for figures, units and missing sections. It takes safety and warning notices exclusively from approved modules. Approval stays with technical writing.
Texts and content at scaleService -
Getting service reports from the field into the system
Service app or form, ERP, CRM
Service technicians write reports on paper, dictate them or type them into their phones in the evening. The workflow turns them into a structured service case: machine, fault pattern, working time, parts fitted, open items. It pre-books the spare parts consumption, prepares the invoice and reports recurring fault patterns to the design department. The technician confirms the report with one click instead of writing it twice.
Tools that fit youService
Systems
Typical systems in mechanical engineering
Your system landscape has grown, often over decades. We work through the interfaces, not the user interface.
- ERPSAP S/4HANA and Business One, proALPHA, abas, Infor, Microsoft Dynamics 365
- CAD and PLMSolidWorks, Siemens NX, Creo, Inventor, PDM and PLM systems with bills of materials and drawing revisions
- CRM and salesSalesforce, Microsoft Dynamics 365 Sales, CAS genesisWorld, product configurators
- ServiceService portals, spare parts catalogues, service apps for technicians, ticket systems
- DocumentationContent management systems such as SCHEMA ST4, translation memories, terminology databases
- CommunicationMicrosoft 365, Exchange, Teams, SharePoint
Objections
What matters most in mechanical engineering
The objections raised first in mechanical engineering companies, and our answers.
- Export control. Technical data on listed goods may require a licence under the Dual-Use Regulation (EU) 2021/821, even if it is only transferred to a server outside the EU. That is why workflows with such data run on models on your own servers or in your EU data centre, agreed with your export control officer.
- Confidential designs. Drawings and design data leave your company only if you decide so in writing for each workflow, and no model is trained on your data; that is written in the contract, not in a brochure.
- Product liability for technical texts. The workflow never writes safety and warning notices itself, it takes them from approved modules, and every version goes through technical writing before publication.
- Works council in production. Our workflows process documents, enquiries and master data, not employee performance; what is logged is fixed in advance, and we supply the documents for a works agreement under section 87(1) no. 6 of the German Works Constitution Act (BetrVG).
- Data that has grown over the years. An item master built up over twenty years is not an obstacle but usually the first workflow: putting data in order is a service of its own with its own fixed price.
Evidence
A case with the same data situation
We do not yet have a mechanical engineering company as a client, and we say so up front. What we do have: two industrial clients with ERP where our workflows have been running for years. At a listed fuel cell manufacturer, Salesforce, SAP SuccessFactors and Freshdesk are connected to one platform with ten domains. At a technical distributor with twelve locations, enquiries land in NetSuite based on rules. The case with the largest data volume comes from the trading sector: a dealer in original industrial spare parts with around 50 employees keeps millions of parts from 2,226 manufacturers in its CRM. The data situation resembles that of a machine builder with a large spare parts business: part numbers in many spellings, placeholders, discontinued products and a system that allows only a limited number of queries.
We cleaned up the data source first, then generated manufacturer texts anchored in facts from the real product range, with eleven quality checks and a rollback that restores any previous state. Result: 3.3 million pages, 71 percent fewer queries to the CRM. The case study also shows what went wrong along the way and what we learned from it.
Prices
Prices
Analysis, build and ongoing operation have fixed prices that are published on our website, and after each stage you decide again. The cost of a workflow such as answering spare parts enquiries or cleaning up the item master is set as a fixed price after the analysis: Approach and prices.
Handover
The first step is a 30-minute call.
You tell us about the workflow that costs you the most time. We tell you honestly whether AI pays off there and what the next step would be. Whether a workflow analysis follows is up to you.