HumanITy Logo
ENDE
Kevin Welter bei der German Speakers Association

Artificial Intelligence

The question is not: How do I use AI? But: What problem do I want to solve?

Many companies start with the technology.

That's the wrong starting point.

Anyone wanting to use AI must first understand what business problem actually needs to be solved. Only then can you assess whether AI is even the right answer.

Often, it isn't.

AI is a tool. Not an end in itself. And certainly not a strategy.

AI is not just AI

In companies, we encounter different forms of AI – with completely different use cases.

01

Classical Machine Learning Models

Algorithms trained on data to:

  • Recognize patterns
  • Make predictions
  • Support decisions

These systems have existed for many years.

They are precise, specialized, and clearly defined.

02

Generative AI

Systems that can create content, interpret, and respond flexibly:

  • Understand and formulate texts
  • Structure information
  • Combine contexts in new ways

In process automation, generative AI is often a powerful tool –

but rarely the foundation.

03

Cognitive AI

Systems that understand complex relationships, interpret, and support decisions.

For example in:

  • Document processing
  • Decision support
  • Structured knowledge management

This is less about creativity –

and more about structural understanding.

The Crucial Point

In many mid-sized companies, the greatest potential lies not in AI.

But in:

  • cleanly defined processes
  • clear architecture
  • classical, deterministic automation

Many workflows can be mapped:

quicklyreliablytraceably

programmatically – without requiring an AI model.

AI in process automation is often the cherry on top of the cake.

Not the frosting. And certainly not the foundation.

When AI Actually Makes Sense

AI delivers value where clearly defined problems can be solved with the right tools.

Data volumes are too large for manual analysis

Patterns cannot be described deterministically

Interpretations become necessary

Forecasts are strategically relevant

Then deployed:

Machine LearningGenerative AICognitive Systems
KI

Deliberately and consciously deployed – not as an end in itself.

The Approach

Technology is not deployed because it's available. But because it meaningfully solves the problem.

1

Understanding the business problem

2

Reviewing process architecture

3

Clarifying what can be solved deterministically

4

Only then deciding if AI adds value

No AI hype.

No tool demonstrations.

But structured decision-making.

Kevin Welter

My Experience with AI Systems

Over the past years, I have practically implemented various AI approaches – always focused on real value rather than technical gimmicks.

📈

Machine Learning

Classical ML models for forecasting and pattern recognition in business processes

✨

Generative AI

Integration into automation processes for text processing and content

🧠

Cognitive Systems

EMMA for intelligent process automation

⚙️

Hybrid Solutions

Combination of AI and deterministic process logic for maximum stability

What matters is not the technology itself –

but its embedding in a sustainable system architecture.

The Goal

🎯

Not "more AI".

But better decisions.

⚖️

Not maximum automation.

But minimum complexity.

💡

Not hype.

But clarity.

Articles

Everything on local AI in one place: hardware, setup, models and data protection.

30 matching articles

Artificial Intelligence for SMEs

Consciously deployed. Architecturally embedded. Business-minded.

Get in touch