---
title: "Punktexe: We advise and deliver"
url: https://punktexe.de/en/
language: en
source: generated from 02_website/content/seiten/en.html
---

**Do It For You**

# We advise and deliver.

Knowing about AI is cheap. Implementation is the challenge. Punktexe implements AI applications for office and industry, such as knowledge assistants built on your own documents, voice agents or the analysis of machine data. We plan, integrate and operate local and hybrid solutions and hand them over to your team on request.

Get in touch, no obligation → How we work

> punktexe --do-it-for-you

![Gentleman lizard in a suit, PUNKTEXE mascot](https://punktexe.de/wp-content/uploads/2026/11/punktexe-echse-440.webp)

**01 · The situation**

## Knowledge is rarely the problem. Implementation often is.

There are plenty of strategy papers, workshops and pilot projects. What is missing are people and structures that get AI into live operation and keep it there.

**Plans without doers**

The idea is there, and so is the pilot. The application never reaches operation because nobody internally has time for it.

**Data that must not leave**

The most valuable data is the most sensitive. It does not belong in third-party clouds without scrutiny.

**Dependent on one provider**

One model, one contract, one price. If you cannot switch, you cannot negotiate, and you come to a standstill when the service fails.

That is why we do it differently.We advise, and then we implement. Like a good concierge: you hand over a task, we know the house and the right people, and we take care of it in confidence and in your interest.

**02 · What we implement**

## We take care of the implementation.

Under one roof: two focus areas and further fields in which we get AI into live operation. What we do not build ourselves is delivered by experienced partners, and the responsibility stays with us.

**01 · Focus**

### Local and hybrid AI

Local models on your own hardware, combined with EU and frontier models where it fits. We select models and hardware, set up the routing and put the application into live operation.

- Knowledge assistant built on your own documents

- Analysis and reporting

- Specialist models per department

Hardware via partners on request, selected per application.

**02 · Focus**

### Industrial AI

From small-batch production to robotics. We assess on site where AI adds value in production and maintenance, which data, models and hardware are needed, and implement the application together with your experts.

- Predictive maintenance

- Analysing machine and quality data

- Assistance for work at the machine

**03**

### Voice agents and communication

We develop voice and telephone agents ourselves and connect them to your systems. They take in requests, classify them and pass them on. Operating location, data flows and running costs are agreed per application.

- Call handling with handover to a person

- Integration via existing user interfaces too, where technically and contractually possible

- Analysis to the agreed and permissible extent

**04**

### Sales and enquiries

Less manual work between enquiry and order. AI does the groundwork, your team decides.

- Following up leads and enquiries

- Qualifying and forwarding enquiries

- Draft proposals from your own building blocks

**05**

### Procurement

Automating recurring procurement steps and preparing decisions better.

- Comparing offers

- Assessing suppliers

- Preparing order proposals

**06**

### Training and enablement

AI changes work. We train teams in a practical way, bring people along and support them in new tasks. In the long term, it is Anna's aim to make AI affordable for social organisations as well.

**07**

### A network that delivers

Development, data, hardware, integration: for each project we select suitable implementation partners from our network. Punktexe coordinates delivery and remains your point of contact.

**08**

### Traceable from the start

Data protection, the EU AI Act and documentation are part of every implementation from the beginning, not proof added at the end.

**03 · Applied AI**

## Your company from above

A factory site with office, production hall, warehouse, access road and energy. Take the tour to see where AI can support the work, what is already in use at Punktexe, what is feasible and where it could develop.

[

![Floor plan of a factory site at night with office building, production hall, warehouse and energy. Illuminated rooms show where AI is already in use at Punktexe.](https://punktexe.de/wp-content/uploads/2026/11/grundriss-nachtansicht.webp)

](https://punktexe.de/en/applied-ai/) [Start the tour →](https://punktexe.de/en/applied-ai/)

**04 · The benchmark**

## What every company should be able to do with AI today

Six capabilities against which you can measure how far your company has come. These are exactly the ones we build up with you.

[Where do you stand? Take the AI check in 4 minutes →](https://punktexe.de/en/ai-check/)

**01 · Use your own data**

Letting AI work with your own data without handing it over.

**02 · Switch models**

Being able to replace the model without rebuilding the application.

**03 · Know the costs**

Knowing what an AI request costs and setting budgets per application.

**04 · Control rights**

Defining who may use which model and what a model may access.

**05 · Be live**

Having at least one AI application in live operation, not just in a pilot.

**06 · Enable people**

Training employees to use AI safely. Article 4 of the EU AI Act requires organisations to take measures towards a sufficient level of AI literacy.

**05 · How we work**

## You hand it over. We take care of it.

Five steps from the first conversation to the running application. You decide where to start.

**01**

### Hand over

You describe the task. A conversation is enough. Documents are welcome but not a prerequisite.

**02**

### Scope

Goal, data, models, hardware and effort. Realistic and in writing, before anything is built.

**03**

### Build

We implement with our own building blocks and experienced partners. You see the progress throughout.

**04**

### Operate

We operate and maintain the application and report regularly on what it achieves.

**05**

### Transfer

When you are ready, your team takes over. On request, not as an obligation.

### The concierge principle

You entrust a good concierge with what matters most: your luggage, sometimes even your children. That trust does not come from status or strategy papers. It comes from knowing the house, the guests and the surroundings, reacting quickly and handling every task in complete confidence and in the guest's interest.

That is how we work. Anyone who wants to improve a company's crown jewels, the processes that have run for decades, has to be on site, know the people and build trust. A visit from outside and a new strategy will not achieve that. We are there, we implement, and we stand by what we do.

That is why we plan with only a few projects at a time, and why discretion and integrity are a matter of course for us.

**06 · Engine room · Architecture principle**

## The shell model: the more sensitive, the further inside.

Our shell model supports the separation of data and model access. The more confidential a piece of information, the further inside it is processed, at the core with local models. Which information stays local and which external services may access it is defined per application. If the classification is unclear, no external processing is intended.

Across all shells we define who may use which model, what it may access and which budget it has. We build our own operations on this principle and plan applications for your company accordingly.

[View the engine room →](https://punktexe.de/en/engine-room/) In the engine room, every status is labelled honestly.

shell-model.exe

**Shell 3 · Frontier models**

**Shell 2 · EU and specialist models**

**Shell 1 · Tools, APIs, research**

**Core · local models**

confidential data, processed locally outward only after approval per application

Rights per shell

Budget per model

Separation per area

Logging and audit

**07 · Where we come from**

## Grown out of practice

From 2016 to today: the professional path of Günther Kriele and his former companies, with overlapping phases, often as side projects and with research and development together with universities and specialists. These stages are not Punktexe client references.

- Recruiting and software for HR processesBuilding a specialised network

- Staffing services for technical industriesData protection in sensitive HR processes

- Automating HR processes with machine learning and AIAutomating HR processes with data protection requirements

- AI consultingLocal, hybrid and cloud models compared

- Today: advising and deliveringPutting AI projects into live operation

[More about us →](https://punktexe.de/en/about/)

**Frequently asked questions**

## Frequently asked questions

### What does “Do It For You” mean?

We do not just advise, we implement. You hand over a task, we scope it, build the solution, operate it and hand it over to your team on request. The burden of implementation does not stay with you.

### How is Punktexe different from traditional consulting?

Traditional consulting often ends with a concept. That is where our work begins: we put the application into live operation and remain responsible for it.

### Can AI run locally on our premises, without a data centre?

For many applications, yes, depending on the size of the model. There are compact AI systems that work without a data centre. We agree in advance which system suits your application.

### Does sensitive data have to go into third-party clouds?

Not necessarily. In the shell model, we define per application which data each model may see. Confidential data is intended for local processing. We only use external models for approved tasks, with data cleansed beforehand where needed.

### Do you also build telephone and voice agents?

Yes. We develop voice and telephone agents ourselves and connect them to existing systems. They take in requests, hand over to a person when needed and are recognisable as AI. We only use automated marketing calls with the prior express consent of the people called.

### Why do you not name any clients?

For reasons of discretion. AI projects change processes and tasks, which is why we only name clients if they expressly wish us to. There are no published references yet for our current focus. In the engine room we show our own developments and label their status.

### How does working together start?

Get in touch with no obligation or send us an email. We will get back to you personally, discuss your project and tell you honestly whether and how we would implement it.

### Who is Punktexe for?

For companies in the DACH region, from mid-sized businesses to large corporations, that want to get AI into live operation, especially with sensitive data and in industry.

**No-obligation enquiry**

## What would you like to implement?

We deliberately take on new projects one after another. Write to us briefly about what it involves: your company or industry, your project and your timeline. We will get back to you personally.

Send an enquiry by email →

We use your details to process your enquiry and will get back to you as soon as we can discuss your project. No newsletter. Information on processing and the service providers we use can be found in our [privacy policy](https://punktexe.de/en/privacy/).
