*Thisarticle is based on a YouTube video by Sophie Hundertmark, an expert in the use of artificial intelligence with a focus on chatbots and strategic AI applications in companies and public institutions. Sophie is a researcher and lecturer at the Lucerne University of Applied Sciences and Arts and is doing her doctorate in Conversational AI at the University of Fribourg. The blog text was created using a custom GPT model that was trained on Sophie’s video content, language style and expertise. The result is well-founded, up-to-date articles based on Sophie Hundertmark‘s own expertise.
You can find the link to the video at the end of this article.
Artificial intelligence is developing rapidly. Many companies have started their first experiments in recent years: a chatbot here, automation there, perhaps an AI-supported analysis tool in reporting.
But this is precisely where a key challenge lies: many of these solutions remain isolated stand-alone solutions.
In conversation with Andrea – an AI expert from the pharmaceutical industry – it becomes clear that the next big step is not to introduce even more individual AI tools. Rather, it is about fundamentally rethinking processes and systematically integrating AI into the entire value chain.
In this article you will find out:
- how AI assistants are already being used in everyday working life today
- what role AI agents play in business processes
- Why “Human in the Loop” remains crucial
- and why companies will have to rebuild their processes in an AI-optimized way in the future
From experiment to the productive use of AI
The AI journey began similarly in many organizations:
With the emergence of generative AI – for example through ChatGPT – the first tests and experiments were launched. Companies were faced with questions such as:
- Do we use external tools?
- Are we developing our own solution?
- How do we handle sensitive company data?
The handling of data is particularly critical in regulated sectors such as the pharmaceutical industry. This is why many organizations decide to build their own internal AI systems.
One example of this is an internal company assistant – comparable to Microsoft Copilot – which supports employees in their day-to-day work.
The internal AI assistant: support in day-to-day work
Such an AI assistant can automate or simplify many typical tasks:
- Coordinate appointments and plan meetings
- Formulating e-mails
- Retrieve information from internal systems
- take on administrative tasks
Seemingly small tasks in particular quickly result in major benefits. A classic example is making appointments with several people – something that otherwise quickly costs a lot of time.
The AI assistant can:
- Analyze calendar
- Identify free time slots
- Book appointments directly
- or formulate e-mail suggestions
This creates digital helpers that take on many small tasks in the background.
Introducing AI in the company: Why role models are more important than duty
One interesting point from the conversation concerns the introduction of such tools in the company.
Instead of a major change management project, the assistant was introduced step by step:
- First simple version of the chatbot
- Continuous improvements
- Integration of internal data
- New functions and releases
Over time, the tool became more and more useful – and its use increased automatically.
What is particularly important here is that managers visibly use the tools themselves.
This ensures acceptance within the team. Instead of preaching “You have to use AI”, simply show how helpful it can be in everyday life.
AI agents: The next step in automation
While AI assistants primarily provide support, AI agents go one step further.
An AI agent is a system that:
- can make decisions independently
- can use different tools
- performs tasks independently
You can think of it like a digital employee who:
- Data analyzed
- Options for action evaluated
- subsequently makes decisions
Of course, this happens within clearly defined rules and boundaries.
Example: AI agents in the supply chain
One practical area of application is supply chain management.
Here, an AI agent analyzes large volumes of data, for example:
- Production data
- Machine breakdowns
- Demand trends
- Stocks
The agent recognizes patterns and trends – for example:
- Machines that break down unusually frequently
- Supply chains that become unstable
- Products with increasing demand
On this basis, he can:
- Issue warnings
- Provide recommendations for action
- or even make decisions automatically in certain cases
For example, an agent can independently make adjustments to replenishment planning – as long as these decisions are within defined risk limits.
Human in the Loop: Why humans remain central
Despite all the automation, one principle remains particularly important: Human in the Loop.
People are actively involved in the process, especially at the beginning of new AI systems.
Typical procedure:
- The AI agent analyzes data and suggests decisions
- A human checks these suggestions
- The system is improved together
- Gradually allowing more automation
This approach reduces risks and at the same time ensures that the system is continuously improved.
This approach is particularly crucial in sensitive sectors such as the pharmaceutical industry.
From individual use cases to an AI operating system
Many companies have now implemented their first AI applications:
- Chatbots
- Automated reports
- Analysis tools
- Small process automations
The problem:
These solutions often exist side by side without being connected to each other.
The future looks different.
Instead of many individual tools, companies need an integrated AI ecosystem – an “AI operating system”, so to speak.
This connects:
- Data
- Processes
- AI models
- Automations
- human decisions
Why companies need to rethink their processes
A crucial point here:
Many company processes have been optimized for decades – or even centuries – for human work.
If you simply integrate AI into these existing processes, the benefits are often limited.
The real leverage only arises when processes are fundamentally redesigned.
This means:
- Making processes AI-compatible
- Thinking automation right from the start
- Viewing humans and AI as a joint system
Only then does the full potential emerge.
Conclusion: The future belongs to integrated AI systems
The last few years have mainly been a phase of experimentation.
Many companies have gained valuable experience – even if not every project has immediately brought great added value.
The next step is clear:
- Connecting isolated solutions
- Rethinking processes
- Strategically integrating AI
Because only when humans and AI really work together can great added value be created.
And perhaps this collaboration will ultimately ensure that we have more time again for the things that are really important – for example, having a personal coffee with colleagues.
Any further questions?
Do you have any questions? I am happy to support you, act as a sparring partner and answer your questions. I am always happy to receive your messages, preferably by WhatsApp message or e-mail.
