AI that works in your company’s daily routine
I help owners and managing directors of small and mid-sized companies use AI where it saves real time. Easy to understand, safe for your data and without a large-scale project.
What AI actually does for your business
Quotes in a fraction of the time
A request arrives by email with a drawing attached. AI extracts the data and drafts a quote using your prices and standard wording. Your sales team checks it and sends it out.
Company knowledge everyone can find
Manuals, service reports and contracts become searchable with a simple question. New employees find answers on their own instead of interrupting experienced colleagues.
Less manual data entry
Delivery notes and incoming invoices are read automatically and passed to your ERP system. Your team only handles the cases that really stand out.
Faster replies to customers
Incoming requests are sorted and replies are drafted, in other languages too if needed. The final decision stays with your people.
Get more out of the AI you already use
Many teams have tried ChatGPT or Copilot and been disappointed. Usually the way of working is the problem, and that can be changed quickly.
Context instead of a single sentence
Before: “Write a reply to the customer.”
Better: Include the complaint, an excerpt from the contract, the tone you want and how far you are willing to accommodate. The result needs hardly any rework.
Team assistants instead of individual chats
Before: Everyone writes their own prompts from scratch.
Better: A shared quoting assistant knows your products, templates and wording. Quality no longer depends on one person.
Connect instead of copy
Before: Text is copied and pasted back and forth between programs.
Better: AI accesses your inbox, file storage or CRM directly, with defined permissions.
Review with a system
Before: Results are either taken over unchecked or not used at all out of distrust.
Better: AI cites its sources, and a person always approves before anything is sent.
Before and after: three typical cases
This is how work changes in companies of your size.
Note: These are example scenarios, not customer references. All times are assumed values for illustration.
Mechanical engineering, 40 employees
Before: Preparing a quote from a customer request takes about 45 minutes. The information is scattered across emails, drawings and old quotes.
After: AI delivers a draft within minutes. Sales reviews it in about 10 minutes and can answer more requests on the same day.
Skilled trades, 15 employees
Before: In the evening, the site manager spends about an hour typing up reports and measurements from handwritten notes.
After: He dictates into his phone on site. AI turns this into a structured report that reaches the office after a quick check.
Wholesale, 25 employees
Before: Two people type orders from PDF files into the inventory management system every day. This leads to errors and follow-up questions.
After: Orders are recognized and entered automatically. The team only handles unclear items.
AI then and AI now
If you checked a few years ago whether AI pays off, it is worth running the numbers again. The barrier to entry has dropped sharply.
Then
- Custom models had to be trained on large amounts of data.
- Projects took many months and required data science specialists.
- Each system could do exactly one task.
- Operation only through specialized software.
Now
- Ready-made language models understand text, images, spreadsheets and speech.
- A first pilot is often up and running within a few weeks.
- AI agents carry out work steps across several programs, and people approve them.
- You use them in plain language, without programming skills.
Your data stays under your control
Sensitive content and AI are not mutually exclusive if the rules are set in advance. This is my approach:
Classify first, then use
We define which data is public, internal or confidential, and which tool is allowed for each level.
The right tool for each level of protection
For many tasks, business plans with a data processing agreement, data storage in the EU and no use of your data for training are sufficient. We check these points for each provider. Particularly sensitive content can stay on a local model on your own premises.
Existing access rights still apply
Document search only shows each person what they are already allowed to open today. Payroll lists stay with HR.
People decide, training provides safeguards
Important steps are approved and logged in a traceable way. Since February 2025, the EU AI Act has also required sufficient AI literacy from everyone who uses AI in a company. That is why the right training is part of the package.
I clarify legal questions together with your data protection officer. My consulting does not replace legal advice.
Questions decision-makers often ask me
Which task takes up most of your team’s time?
Tell me about it. In the initial consultation, we check whether AI really helps there and what the most sensible first step is.
