When processes are repetitive, data-rich, or critical for decision-making, e.g., for automation, analysis, or personalization.

What is the best way to get started with AI?

It is better to start with small, clearly defined use cases rather than tackling entire projects right away.

What are the risks associated with using AI?

Data quality is crucial. Poor or unstructured data leads to unusable models. Clear processes, responsibilities, and responsible use are also required.

What problems in our company can AI solve?

AI is particularly suitable for tasks such as pattern recognition, predictions, automation, or intelligent evaluation of data. We will work with you to determine whether your use case is suitable.

How quickly will we see results?

We often start with a proof of value. This gives you an initial assessment within a few weeks of whether and how AI can support your goal.

What costs do we need to budget for AI implementation?

That depends heavily on the use case and the data. It is important that a project delivers a clear business case and measurable results.

What does AI mean for our employees?

AI does not replace skilled workers, but supports them. Routine tasks are completed faster, leaving more room for value-adding activities.

Services

Digitalisation Artificial Intelligence

Artificial intelligence: When should your company use AI?

How can I support you?

Samuel Picek
Lead ML/AI Engineer

+41.. Show number +41 55 253 00 53

Get expert advice. Discover our success stories.

Artificial intelligence enables companies to automate inefficient routines, improve data-based decisions, and optimize strategic processes in a targeted manner.

This is why artificial intelligence makes sense for companies:

Increased efficiency through automation: AI takes over tedious, repetitive tasks. This leaves time for value-adding activities.
Data-driven insights instead of gut feelings: AI models recognize patterns in data and deliver more accurate analyses, forecasts, and recommendations for action.
Customized solutions for real challenges: Whether chatbots, predictive maintenance, or automated quality control, AI helps with specific use cases.

Frequently asked questions

When is AI useful?

When processes are repetitive, data-rich, or critical for decision-making, e.g., for automation, analysis, or personalization.
What is the best way to get started with AI?

It is better to start with small, clearly defined use cases rather than tackling entire projects right away.
What are the risks associated with using AI?

Data quality is crucial. Poor or unstructured data leads to unusable models. Clear processes, responsibilities, and responsible use are also required.
What problems in our company can AI solve?

AI is particularly suitable for tasks such as pattern recognition, predictions, automation, or intelligent evaluation of data. We will work with you to determine whether your use case is suitable.
How quickly will we see results?

We often start with a proof of value. This gives you an initial assessment within a few weeks of whether and how AI can support your goal.
What costs do we need to budget for AI implementation?

That depends heavily on the use case and the data. It is important that a project delivers a clear business case and measurable results.
What does AI mean for our employees?

AI does not replace skilled workers, but supports them. Routine tasks are completed faster, leaving more room for value-adding activities.

Why do companies use artificial intelligence?

Companies use artificial intelligence to increase their efficiency and automate complex tasks. By using artificial intelligence in software or, for example, IoT systems, complex problems can be simplified or even completely automated. Repetitive or time-consuming tasks can be taken over by AI, for example to free up resources for more effective activities.

AI as part of software projects: our experience

Artificial intelligence is not always the better solution. We use it where traditional approaches reach their limits and AI brings real added value. An example from autonomous driving illustrates this well:

Detection (AI): Data from radar, cameras, or lidar is highly unstructured. Classic algorithms can hardly reliably detect cars, bicycles, or pedestrians. AI, on the other hand, recognizes patterns in this data and delivers stable results.
Tracking (classic algorithms): Once objects are detected, their movements can be described using known physical laws. Here, proven methods such as Kalman or particle filters work better than AI.

This illustrates that AI plays to its strengths wherever huge amounts of data and unclear patterns are involved. Where rules are clear, classic software remains superior.

This means that AI only makes sense if it delivers a noticeable improvement, whether in terms of result quality or development time. In highly regulated areas or deterministic processes, we continue to rely on classic software.

When does artificial intelligence makes sense?

Use it when	Avoid it when
classical algorithms reach their limits klassische Algorithmen an ihre Grenzen stossen	deterministic processes with clear rules dominate deterministische Prozesse mit klaren Regeln dominieren
large, unstructured amounts of data are available grosse, unstrukturierte Datenmengen vorliegen	highly regulated tasks can be solved without AI stark regulierte Aufgaben auch ohne KI lösbar sind
measurable improvements in quality or development time are possible messbare Verbesserungen bei Qualität oder Entwicklungszeit möglich sind	compliance and ethical risks make implementation difficult Compliance- und Ethik-Risiken die Einführung erschweren

Differences between traditional software and AI

Traditional software is based on clear rules. Developers specify step by step how the system should respond. As long as the inputs are correct, the same predictable result is always produced.

AI works differently. Instead of programming fixed rules, we train models with large amounts of data. These models recognize patterns and probabilities. The results are often very accurate, but not always 100 percent predictable. This means:

Traditional software: when rules are clear and deterministic, e.g., calculations, business logic, or processes with fixed procedures.
AI software: when patterns, probabilities, and unstructured data are paramount, e.g., image recognition, language processing, or predictions.

Project differences at a glance

1. Classic software

Clear requirements and acceptance tests
Stable range of functions over many years
Less dependence on data quality

2. AI projects

Exploratory approach with experiments
Data preparation and quality are crucial
Success measured using benchmarks instead of fixed tests

AI projects are more dynamic and data-driven. They require iterative development and close collaboration between business and technology.

Success factors for AI projects

Data quality and data volume

AI needs good data. If the data is incomplete, incorrect, or contradictory, even the best model will not deliver usable results.
Clear goals and benchmarks

An AI project must be measured against specific questions: What should the solution be able to do better than before? Which key figures indicate success? Without clear benchmarks, the result remains vague.
Iterative approach

AI development rarely follows a rigid plan. Instead, we test hypotheses, adapt models, and improve them step by step. Small pilot projects quickly show whether an approach works.
Collaboration between business and technology

AI can only create added value if business and IT work closely together. Business departments know the processes and challenges, while developers contribute the technical expertise.

Sam Picek, Lead AI/ML Engineer, soxes AG

What AI solutions does soxes offer for businesses?

Welche KI-Lösungen bietet soxes für Unternehmen an?

At soxes, we develop customized AI solutions for businesses that are tailored to their specific needs. Our AI development includes, among other things:

Data analysis and predictive modeling: Using AI to analyze large amounts of data and create predictive models that help businesses make informed decisions.
Business process automation: Using AI to automate repetitive tasks, thereby increasing efficiency and productivity.
Personalized customer interaction: Developing AI-powered chatbots and recommendation systems that improve and personalize customer interaction.

Here you will find an overview of some of the projects we have already implemented with AI!

Use case: Prototype for error detection at Emmi

1 Project Overview

Emmi, one of the world's leading manufacturers of dairy products, uses state-of-the-art technology in its production facilities. soxes developed an AI prototype for Emmi for automated error detection in the yogurt production process.
2 Challenge

Yogurt packaging must be flawless so that products can be correctly labeled and further processed. The goal was to reliably detect deviations in real time.
3 Solution

The AI application processes images of packaging captured by cameras. Using Microsoft Custom Vision and a Python app, the images are analyzed and compared with product data. Deviations, such as an incorrect lid, are immediately detected and reported.
4 Result

Within six weeks, an efficient solution was developed that ensures production quality in real time. Faulty products are identified at an early stage, reducing waste and increasing process reliability.

How we use AI in software development

Wie wir KI in der Softwareentwicklung nutzen

We use AI not only for customer projects, but also in our daily development work. We have gained a variety of experiences in this area:

Coding assistants

Tools such as GitHub Copilot and JetBrains AI help us write code. They suggest solutions, generate tests, take care of boilerplate code, and assist with debugging. This saves time and allows us to move forward more quickly.

Coding agents

Platforms such as Replit or GPT-based Codex are very good at implementing small prototypes. However, they still have their limitations when it comes to large code bases or complex refactoring.

Team Impact

AI is changing how developers work. Routine tasks are being automated, leaving more room for conceptual work and problem solving. Today, generalists with a good overview are more in demand than pure specialists for repetitive tasks.

Comparison with industry

We often compare this to CNC machines in metalworking: there, too, the machine does not replace expertise, but rather relieves the burden of repetitive work. Without a solid foundation, neither CNC nor AI can be used effectively.

For us, AI is a tool that speeds up work and improves quality. However, the expertise of the team always remains crucial.

How much AI potential does your company have?

Unlock your AI potential with our consulting services!

soxes offers ONE-to-ONE consulting tailored to your situation, focusing on the individual opportunities for your company.

We show you exciting possibilities and helpful solutions. Simple and competent.

Go to Article

3D illustration of a row of friendly white AI robots working on blue laptops at small desks, symbolizing artificial intelligence consulting for business.

Our approach to the responsible use of AI

Unser Ansatz für den verantwortungsvollen Einsatz von KI

AI is a powerful tool. For it to deliver real benefits, clear guidelines are needed:

Transparency

We only use AI where we can understand how it works. Decisions must remain explainable to customers, users, and employees.

Data protection and security

The protection of sensitive data is a top priority. We ensure that data is processed responsibly and stored securely.

Ethics and fairness

AI must not reinforce existing prejudices. We check models for possible biases and ensure that results remain fair and balanced.

Focus on people

AI does not replace expertise. It supports people in making better decisions and working more efficiently. Responsibility always remains with humans. Our goal is to use AI in a way that creates added value, minimizes risks, and empowers people rather than replacing them.

Find out if your company is AI-ready

Not every company is immediately ready for AI. We show you which steps make sense and where you can start.

Analysis of data, processes, and goals
Clear roadmap for the next step

Direct number

+41 55 253 00 53

Start your AI check now!

Vier Personen sitzen am Tisch und diskutieren über Künstliche Intelligenz (AI steht auf einem Whiteboard hinter ihnen)

How can we support you with your software project?

Samuel Picek
Lead ML/AI Engineer

+41.. Show number +41 55 253 00 53

Please contact me Our References

Expert knowledge for your project

Sam Picek is an AI/ML engineer at soxes with a master’s degree in mechanical engineering at University of Zagreb. He brings many years of experience in machine learning, data engineering, and software architecture to the table. His strengths lie in Python, Rust, Kafka, CI/CD, cloud platforms, and the development of intelligent systems for complex data landscapes.