Quote and order intake
Data from emails and PDFs goes into the ERP or quote sheet, with similar past jobs attached. A person always sets the price.
Practical AI for manufacturers: an AI workspace, team training and paperwork automation.
Your staff learn to use AI safely on your own requests for quote, orders and manuals, and the hours spent retyping come back.
- On site in Poland and the EU
- Fixed price, no platform licence
- NDA before we see any document
- A person approves every output
- Every answer shows its source
08:12
Hi, could you quote 250 pcs of bracket WS-114 in S355, 8 mm plate, bent and galvanised? We’d need delivery by 14 Nov. Thanks, Jan
The quote sheet fills itself in
Quote 2026-118
- 1Item
- WS-114 bracket × 250
- 2Material
- S355, 8 mm
- 3Process
- bending, hot-dip galvanising
- 4Delivery
- 14 Nov 2026
- Similar job
- WS-108, 300 pcs, Mar 2026: €6.50/pc
- Price / pc
- entered by your estimator
Where the hours go today
Quotes wait for days.
Requests for quote (RFQs) arrive by email with drawings attached. Someone retypes them into the quoting sheet by hand before any real estimating starts.
Orders are retyped from PDFs.
Line by line, from the customer's PDF into your ERP system. Every typo turns up later on the shop floor.
Know-how lives in two people's heads.
How to set up a job, where the procedure is, what went wrong last time: your technologist knows, and maybe one other person. The rest sits in folders nobody can search.
Company files go into free ChatGPT.
Your people already use AI on their own. Nobody has said what's fine to paste in and what isn't.
It's hours every week, and hiring another person for the paperwork usually isn't an option.
How we work together
Three steps. The first is completely free.
- Step 1 · freeWorkflow check45 minutes online, then a written audit report.
- Step 2AI KickstartTeam training and a set-up AI workspace.
- Step 3 · optionalAutomationsWe build what the Kickstart shows is worth it.
Step 1
Free workflow check
A 45-minute conversation online. Within a few days you get a written audit report covering:
- how your team uses AI today,
- where paperwork takes the most time,
- where AI can help, and where simpler solutions will do,
- our honest view on whether the Kickstart makes sense for you.
Step 2
AI Workspace Kickstart
Training that gets your team started
A few days with your key people, on site or online. We train your team and set up a professional AI workspace fitted to the way you work.
How it runs
The week before
Preparation
A short call, an NDA, and agreeing the scope of the workshops.
Training
Tools and good practice
Today’s tools, what they do well and where they fall short, and how to use them effectively and safely.
Workspace
Set-up and integration
We connect your data sources and set up an AI workspace that works with your company’s context.
Workshop
Build together
Each person builds 1–2 helpers for their own tasks. Together we write down where work gets stuck.
What you keep
- A business AI workspace that knows your company’s context
- Training materials and rules for using AI
- Tested prompts and assistants that help with everyday work
- The skills to extend the workspace with further features
- A list of promising directions for further automation, with rough hours saved
Step 3 · optional
What comes next: example automations
Once the Kickstart shows where your team loses the most time, we can automate it. Below are three common directions, but the scope always comes from what we find in your company. Fixed price, agreed after scoping.
Shop-floor knowledge assistant
Answers from manuals, procedures and non-conformance reports, always with the document and page. Can run as a dedicated cloud deployment or on your own server.
Reports from data you already have
We automate the 3–5 reports people rebuild by hand from ERP and Excel exports.
What it looks like in practice
Visualisations of example automations
Quote intake
Knowledge assistant
Reports
Questions
Is our data, and our customers’ drawings, safe?
Business-tier tools don’t train models on your data, and sensitive documents can stay on your own server. We sign an NDA before we see any document.
Drawings under confidentiality only go to a standard cloud AI service if the customer contract allows it. Otherwise we use a dedicated cloud deployment, set up for your company alone, or a model running locally, depending on what the contract allows.
Which tools do you use? Do we need licences or an IT department?
Usually the business version of a mainstream AI assistant, and where data is sensitive, a dedicated cloud deployment or a model running locally. You buy licences directly from the vendor; we don’t sell a platform of our own.
You don’t need an IT department. We do the setup.
How soon will we see results? And what if the team won’t use it?
The first ones during the training: each person leaves with 1–2 helpers for their own tasks. We practise on the team’s real work, and after the training we offer follow-up sessions to fix whatever isn’t being used.
Does the training meet the EU AI Act article 4 requirement?
The Kickstart supports it: the training is documented (programme, attendance record, summary) and fitted to how your team uses AI. This isn’t legal advice.
Do we have to buy an automation after the Kickstart?
No. You keep the tools, the rules and the list of automation opportunities. If something on that list is worth building, we’ll quote it at a fixed price.
Book your free workflow check
45 minutes online, an audit report afterwards. Free, no obligation.