AI and software for manufacturers
We apply AI in manufacturing and do manufacturing software development for discrete and process manufacturers and suppliers in the high-tech supply chain. We connect ERP, MES and machine data, and build the quoting, quality and maintenance tools your engineers and planners decide with.
Where the work gets stuck.
Quotes wait on engineering
Make-to-order suppliers receive RFQs with drawings, STEP files and specifications. Engineers estimate routing, material and hours by hand, and slow answers lose orders.
Machine data stays on the machine
PLCs and controllers produce useful signals, but they rarely reach ERP or planning. OEE, scrap and downtime are recorded on paper or typed in at the end of a shift.
Quality knowledge is buried in reports
Non-conformance reports, 8D analyses and customer complaints pile up in PDFs and the ERP, so the same root causes are investigated again on a different line.
How we improve manufacturing.
Each one is scoped around your systems and rules, and each one keeps a person in charge of the decisions that matter.
RFQ intake and quote preparation
RFQ emails, drawings and specifications are read into a structured request, matched to similar past orders and prepared with a suggested routing and material list. The estimator sets the price and approves the quote.
Shop floor data pipeline
Machine signals collected over OPC UA and MQTT, combined with orders from MES and ERP into one production data model. Production managers see OEE, scrap and downtime per order and shift.
Quality and deviation search
Non-conformance reports, 8D analyses, test results and customer complaints searchable by meaning, so a quality engineer finds similar past issues and the fixes that worked. The engineer decides the corrective action.
Maintenance technician copilot
Technicians ask about fault codes, procedures and past repairs and get answers from machine manuals and maintenance history, with sources. Lockout and safety procedures are shown as written, and the technician follows them on site.
ERP and MES integration
Production orders, material bookings and hour registrations synchronised between ERP systems such as Isah, Exact or SAP and the MES, with errors in a monitored queue. Planners no longer enter the same order twice.
Dealer and spare parts portal
A portal where customers and dealers find parts from exploded views and serial numbers, check stock and order, fed from PLM and ERP data. Service engineers handle the requests the portal cannot answer.
Built around the rules.
What we design for from the first week. Your legal and compliance people keep the final word.
NIS2 for manufacturers of critical products
Manufacturers of machinery, vehicles, electronics and medical devices can fall under NIS2 as important entities. We design for access control, logging and supply chain security and document it for your assessments.
IEC 62443 for OT connections
Machine data is collected through segmented, read-only connections aligned with IEC 62443, agreed with your OT team. Nothing we build writes to PLCs or safety systems.
Machinery Regulation and the AI Act
The EU Machinery Regulation applies from January 2027 and puts safety components with machine learning behaviour in its highest-risk category. Our software supports people and processes around the machine and stays out of safety functions.
EU Data Act for connected products
Users of connected machines have rights to the data their equipment generates. We structure product and usage data so it can be shared with customers and third parties on request.
Works with what you run.
If a system has an API, a database, an export or an inbox, we can build on it. These are the ones we meet most.
- SAP S/4HANA
- Microsoft Dynamics 365
- Exact
- Isah
- Ridder iQ
- MES platforms
- OPC UA
- MQTT
- Siemens Teamcenter
Where to start.
RFQ intake for one product family
Quoting is document work with a measurable outcome: time to quote. We start with one product family and your estimators, and test the output against orders you have already quoted.
Talk it throughWhat it includes
- Extraction of RFQ emails, drawings and specifications
- Matching to similar past orders in your ERP
- Draft routing and material list for review
- Evaluation against recent quotes
Guides.
AI readiness checklist: what to answer before your first agent goes live
The questions to answer before your first AI agent handles real work, from process and data to approvals, monitoring and ownership.
10 min read
Build vs buy software: a decision guide for operations and product teams
A practical guide to choosing between off-the-shelf and custom software, with a cost comparison, a decision matrix and a checklist for your team.
10 min read
Further reading.
Embeddings and semantic search: a practical guide for product teams
How embeddings power semantic search, how to choose a model and a vector store, and how to measure whether your results actually improved.
5 min read
Document parsing with LLMs: extracting data from invoices and forms
How to turn invoices and forms into reliable structured data with LLMs, schemas, validation and a review queue for the cases that need a person.
4 min read
RAG vs fine-tuning: which does your business actually need?
RAG gives a model your knowledge at answer time; fine-tuning shapes its behaviour. How to choose, when to combine them and what each costs.
4 min read
Common questions.
Yes, if the controllers or an edge gateway expose OPC UA, MQTT or a database. For older machines we work with your automation partner to add a gateway, and we always read data without writing to the controller.
Yes. Both have APIs or data access we can build on, and we also work with SAP S/4HANA and Dynamics 365. We confirm the interfaces with your ERP partner during discovery.
Your OT team sets the rules. We use segmented networks, read-only connections and approved gateways aligned with IEC 62443, and we never need access to the control network itself.
Models handle title blocks, material notes, tolerances in text and parts lists well, and geometry less reliably. For quoting we combine what the model reads with your past orders, and the estimator checks everything that affects price.
An RFQ intake build for one product family usually takes weeks. Shop floor data projects depend more on network access and machine types, so we map those before we plan.