Working smart with AI: The art of prompt engineering

Over the past few years, artificial intelligence has captured our imagination. Generative AI in particular stands out, because we can see it at work in our daily lives and jobs. We ask a tool to write a text, create an image, or even generate a short video. Where we once needed specialized software and technical expertise, today a clear instruction in plain language is enough.

That accessibility makes the technology attractive, but also misleading. Anyone who assumes generative AI automatically delivers perfect results will be disappointed. No matter how fast AI evolves, its output can still be mediocre or flat-out wrong. That’s why prompt engineering matters: the skill of formulating smart questions and instructions.

What is generative AI?

Generative AI is a form of artificial intelligence that creates new content based on existing data: text, images, video, code, or audio.

Its strength lies in the massive datasets the models are trained on. They recognize patterns and use that knowledge to predict the most likely outcome: the next sentence, the most fitting image, or the right piece of code.

In broad terms, generative AI works in four steps:

1. Training. The model learns to recognize patterns from vast datasets.
2. Processing the prompt. The input (prompt) is analyzed and “understood.”
3. Generation. The model predicts and produces a fitting response.
4. Refinement. The output is adjusted and fine-tuned to match the requested parameters.

One key point: generative AI doesn’t “think”. It’s a probabilistic model that calculates the most likely outcomes based on learned patterns. That means it can deliver impressive results, but also produce errors.

Applications in work and daily life

The possibilities are broad. In business, generative AI supports software development, data analysis, customer service, and even cybersecurity. Think of chatbots that handle standard questions or models that detect system vulnerabilities.

At home, the technology is just as useful. Generative AI can help you tell your child a bedtime story with their favorite characters, plan a holiday, or explain a complex subject in simple terms.

Getting better results with prompt engineering

Typing a quick question and expecting the perfect answer… rarely works. Generative AI is powerful, but its true value emerges when you know how to craft effective prompts. Prompt engineering is the art of instructing an AI system so that it clearly understands what you mean and what you need.

What is prompt engineering?

Prompt engineering goes beyond a Google search. It’s about deliberately guiding AI models by:

• clearly stating your intent,
• adding context or examples,
• and refining the output step by step.

The better you are at this, the more useful and reliable the results become.

Common prompt engineering techniques

How you build a prompt depends on your goal. These are the most common techniques:

• Chain of Thought. Break a complex question into smaller logical steps. The AI follows your reasoning and produces structured answers.
• Priming. Add examples, definitions, or context. This helps the model better understand tone, style, and nuance.
• Iterative refinement. Use the model’s output to improve your next prompt. By iterating, you get closer to the desired result.
• Zero-shot prompting. Ask a question without extra context and rely on the model’s general knowledge. Useful for simple tasks.
• Few-shot prompting. Provide a few examples in your prompt, so the model learns the desired style or structure.

Effective vs. less effective prompts

A vague request usually leads to a vague response. Compare:

• “What are the benefits of meditation?” → a shallow list.
• “List the five main benefits of meditation, explain them briefly, and connect them to recent scientific insights.” → concrete, useful, and better supported.

Another effective trick is to assign the AI a role: “You are a software engineer reviewing code” or “You are a teacher explaining this to a class of 12- and 13-year-olds.” This steers the tone and perspective of the output.

Why prompt engineering matters

Good prompts save time, make results more reliable, and let you use AI as a creative sparring partner rather than just a gimmick. Prompt engineering is becoming an essential skill: if you know how to work smart with AI, you stay ahead.

Using generative AI responsibly

Generative AI can deliver great results and save time, if used with care. The technology is powerful, but also fragile. Misuse can lead to misinformation, privacy issues, or even security risks.

What to watch out for

Common pitfalls when working with AI include:

• Inaccuracies and hallucinations. AI often presents information with confidence, even when it’s wrong. Always verify with reliable sources.
• Bias and fairness. Models learn from existing data, including biases. Output can unintentionally be discriminatory.
• Copyright and attribution. Generated content may contain elements of existing material. Proper attribution remains essential.
• Over-reliance. AI is a tool, not a replacement for human judgment. Stay critical.

Safe and responsible use in practice

Whether using AI privately or professionally, safe and responsible use requires clear rules. That means handling data carefully and staying critical of results. In Europe, GDPR plays a central role: organizations must limit data to what is strictly necessary, be transparent about usage, and prove that data is processed responsibly. For companies, there is an extra duty to protect sensitive business information through clear guidelines and secure platforms.

Practical guidelines for safe AI use:

• Anonymize data and never use names or confidential details in prompts.
• Always verify AI output with trusted sources.
• Use reliable solutions that meet security standards.
• Follow your organization’s internal guidelines and processes.
• Keep business and personal data strictly separate.
• Share only the minimum necessary information (data minimization).
• Use data only for its original purpose (purpose limitation).
• Be transparent about when and how data is processed.
• Document policies and conduct impact assessments if needed.
• Train employees so everyone handles AI responsibly and with awareness.

By following these principles, you’ll not only comply with GDPR but also avoid exposing sensitive company data. This way, you reap the benefits of generative AI without unnecessary risks.

The power of the TMC A.I. HUB

AI now plays a role in nearly every field, including all of TMC’s service areas: Technology & Engineering, Digital & IT, Energy & Renewables, and Life Sciences & Pharma. Because generative AI and related technologies are becoming so widespread, we want to pool our knowledge and experiences. That’s why the TMC A.I. HUB was created.

The HUB is a meeting place where colleagues and clients exchange ideas, inspire each other, and develop concrete AI applications. We learn from each other, discover new opportunities, and test them together.

The next step
The A.I. HUB is growing step by step. Training sessions, workshops, and international exchanges will allow us to share experiences. Pilots and projects will also be launched to apply AI directly in practice. Ultimately, the goal is for AI to no longer be seen as a separate activity, but as a natural part of our work and of the value TMC delivers to clients.

Join us
Want to learn more about the TMC A.I. HUB or join this community? Send a message to ai.hub@tmc.nl or reach out via TMC’s internal channels. Together, we can make AI a powerful asset in our daily work, and for our clients.