Geschäftslogik, die funktioniert.
KI-Lösungen · CRM & ERP · Hosting in Deutschland
Engineers of business logic: process analysis, CRM & ERP, AI employees, custom software and industrial AI for companies with 5 to 250 employees. Built from scratch or combined from the systems you already have: we use whatever solves your task. We measure before we start and show you the number afterwards.
Free of charge and without obligation. 30 minutes, one concrete process from your daily work, no slides.
Servers in Germany · AI advises, a human decides · Transparency under the EU AI Act
Try it yourself
Move the pointer over the sheet and click a station. The switch on the right hides everything that isn't running yet: what remains is what we have already built at customers. The difference between the two pictures is not the technology, but the rule it works by.
Picture of the future · who carries it out changes, the rule stays
Note: The drawing shows a workflow, not a catalogue of services. Vehicles, drones and devices stand for whoever carries out the work: haulage companies, trade businesses, manufacturers. They are not products or services of W&K Connect GmbH; we do not manufacture, operate or broker them. Our contribution is solely the software logic: process automation, AI employees and connecting CRM, ERP and existing systems. Dotted parts are a picture of the future and not available today. What this means in detail · What we deliver today
The gap in the market
You know how it goes: a tool is set up, a few instructions are written, the demo looks good. Then your people start working with it, and two weeks later you hear: “This isn't what we need.” You test, send feedback, wait for the next version. And test again.
The reason is rarely the technology. The reason is that nobody understood your business beforehand:
That is the gap. The market confirms it: 45% of the companies surveyed name a lack of staff resources as a problem when using AI, 42% insufficient data quality (survey by VDMA, the German mechanical engineering association, published December 2025). Neither is a technical problem. Both are problems of process and logic.
Source: report on the VDMA survey on AI in mechanical and plant engineering, der-maschinenbau.de (in German), 18 December 2025.
That is why we start with your process, not with software. → How we work
1 VDMA survey, published December 2025.
One partner for the whole journey
Wherever you are now: you get started where your problem is and keep growing with us. You don't need five service providers for that.
Stage: entry
Your systems don't talk to each other. Data is retyped, emails are sorted by hand, quotes are pieced together from three sources. We connect what you already have and take the routine off your team's hands.
InterfacesInboxData reconciliation
AI integration & process automation →Stage: core
You want to run your business in one system: customers, orders, stock, invoices. We implement CRM/ERP along your process or develop software that doesn't yet exist in that form.
Process analysisCRM/ERPAI agentsSoftware
CRM & ERP with AI → Custom software & AI agents →Stage: industry
Your machines produce data, but nobody analyses it. A single unplanned downtime can cost more than a year of maintenance. We lay the logic over your existing equipment: detect the signal, create a maintenance work order in the ERP, measure the effect.
SensorsAnomalyWork order in the ERP
Predictive maintenance & industrial AI →Stage: industry and future
Robots, cameras and machines do their work, but they don't think ahead. We build the logic layer on top: coordinate tasks, detect exceptions, feed data back into the ERP. Safety functions stay with the manufacturer. And in the Future Lab we show what becomes possible next.
RoboticsCamera AIEdge
Being built AI for machines & robotics → Future Lab →AI employees · product
One AI agent per role – from management through purchasing, warehouse, sales and accounting to drivers and the phone – works in your ERP, CRM and telephony. With a memory for your rules, with the permissions of its role, with limits: decisions on money, contracts and staff are made by a human.
RoleMemoryPermissionsApproval
All five services side by side, with the question of which stage suits you: → Services at a glance
How we work
01
We go through the process with your people. Not the one on the organisation chart, but the one that actually happens every day: with exceptions, detours, paper notes. Duration: one or two sessions on site or by video.
02
We write down what we have understood: who decides what. Which data comes from where. Which rule applies in an exception. Where a human has to approve. You read this document and correct it. You sign it off. It is the basis for everything that follows, and the reason why our systems don't end up being “not what we need”. We call it the logic document, and there is no project without one.
03
Now we decide what to build it with: connect your existing systems, implement a standard system or develop something new. The answer follows from the logic, not from our product catalogue. We don't have one.
04
A fixed-scope pilot. First we measure the baseline: time per case, error rate, downtime. Then your employees work with the system for 2–4 weeks. After that, the number is on the table: does it work or not. Your number, not our promise.
05
Rollout to further departments, sites or equipment. Training for your employees, because they are the ones who carry the system. Ongoing operation, if you want it.
Typical starting situations
Goods arrive in the morning and have to be weighed, labelled and out on the route by noon. Between scale, label, inventory management system and route plan there are four manual steps that nobody needs.
What we do here: close the chain, turn the knowledge into rules, give every role an AI employee. → Food & logistics
Buying from private individuals and businesses, several VAT rules depending on the goods, supporting documents for identity and origin. Today all of it depends on your employees' memory.
What we do here: duties as system rules, so that the supporting document doesn't depend on memory. → Trade & precious metals
Dozens of projects at the same time, each in a different phase. Anyone who wants to know the status calls the site manager. The CRM was built for sales, not for construction workflows.
What we do here: the project as one continuous workflow with phases, required documents and the people responsible. → Construction & trades
The IT has grown over the years. Nobody has a complete list of devices, addresses and access credentials. Faults are guessed, not measured.
What we do here: first an inventory, then monitoring, then operation, with a fixed point of contact. → AI integration & process automation
Trust
We work for companies in trade, food, construction, the care sector and manufacturing, from family businesses to mid-sized companies with several sites. We don't name them: discretion is part of our work. Anyone who opens up their goods receipt, their pricing logic or their maintenance to us should be sure that none of it turns up anywhere as a reference. We can arrange reference calls on request if the customer agrees.
A logic document that the customer signs off
Every project starts with a logic document that the customer signs off. No exceptions. It is the reason why our systems are used, not just accepted.
AI employees with memory and permissions
AI employees with their own memory and their own permissions work today in more than 20 roles, from the driver role to management. Each one sees only what its task requires.
Legal duties as system rules
We have built ERP modules in which the supporting document can be set up as a required field: German Anti-Money Laundering Act (GwG), VAT, waste law. The supporting evidence is a report, not a folder.
AI infrastructure on servers in Germany.
We operate memory systems, agent logic and your business data there, kept separate for each customer. Critical data runs offline and locally; we connect external language models only for non-critical data and only as processors under Art. 28 GDPR.
We operate what we build
We work with AI employees ourselves, in management and accounting.
Discretion as a principle
In the first call we tell you what we have built, without saying for whom.
That is enough to judge whether we are the right fit for you. Who we are: → About us
Trust and responsibility
We build AI that makes suggestions, prepares data and handles routine work. Decisions on money, contracts and staff stay with your employees. This is not just our stance: it is built into every system as an approval step. What the permissions, memory and limits of an AI employee look like in detail: → AI employees
Every AI employee we build for you identifies itself as AI: to your employees and, where it communicates externally, to your customers. This obligation has applied since 2 August 2026. We check the risk class of your project before the start, not at acceptance. → AI transparency
Your data is stored on servers in Germany. Each AI employee has only the permissions its role needs: the driver doesn't see purchase prices, purchasing doesn't see salaries. If we process personal data on your behalf, we conclude a data processing agreement (DPA) with you under Art. 28 GDPR; every service provider involved, including the provider of a language model, is listed in it. → Privacy policy
Before every pilot we define how you will recognise success. After the pilot we present the number. If it isn't right, you hear it from us first.
Based in Germany · GmbH · HRB 14751 NP (Neuruppin)
Questions
With a process from your daily work, not with software. In the first call we go through it in 30 minutes; afterwards you receive one page with our assessment.
How the first call works is described on the contact page: → Contact
A logic document describes your process in your own words: who decides what, which data comes from where, which rule applies in an exception, where a human approves. You read it, correct it and sign it off. Only then do we build, which is why the result fits your company.
In our process, this is the second step: → How we work
For companies with 5 to 250 employees that run several systems or have a task that no standard software covers. The smallest pilot is one AI employee for one role. You don't have to be a large corporation.
On three things: the scope of the process, the systems we connect and who operates the result. We give you a price range in the first call, based on your process, not on a price list.
The models and frameworks are on the pricing page: → Pricing & pilot models
If your company has processes, yes. We know trade, food, construction, the care sector and manufacturing from day-to-day operation; the heart of our work is the same everywhere. We offer process mapping, the logic document and training in German, English or Russian.
Which industries we describe in detail: → Industries
Future Lab
We expect that over the next few years robots, cameras and AI systems in small and mid-sized companies will take over tasks that people do today – not because it is modern, but because there aren't enough people. Anyone who waits until then will be too late. We are building the building blocks today: the cycle from sensor to maintenance work order, the logic above the machine, AI colleagues that work inside your systems.
First call
In the first call we take one concrete workflow from your business: the one that costs you the most time or frustration. After 30 minutes you know whether and how it can be solved. No obligation, no slide decks.
Reply within one working day.