IT skills & AI

AI is no longer just a chatbot or a text generator. It's a new layer of work that can help you think, create content, analyze information, write code, use tools, and carry out work steps under human direction. In practice, AI can summarize documents, analyze images and data, retrieve information from company systems, prepare decisions, automate routines, build prototypes, and operate as part of broader workflows.

Even so, today's AI models don't "think" the way humans do. They generate answers and actions based on their training data, the context provided by the user, the tools available to them, and the permissions they've been given. That's why AI can be an extremely effective partner — but it can also make mistakes, use the wrong information, draw faulty conclusions, or act incorrectly if its goals, data, and boundaries are poorly defined.

Understanding AI fundamentals therefore means more than writing good prompts. It means understanding when AI is best used for brainstorming, when for production, when for analysis, when for automation, and when for work done with tools or agents. The more access AI gets to data, systems, and decisions, the more important it is to clearly define permissions, oversight, human approval, and accountability.

The layers of AI describe how modern AI is built, from energy, data centers, and chips all the way up to language models, data, integrations, agents, applications, business workflows, and management. AI isn't just a single chatbot or app but an entire value chain, where every layer affects what AI can actually do.

AI only creates value when the layers work together. A powerful language model needs the right data and context. Data needs secure access controls. Agents need integrations, tools, and orchestration. Companies need clear processes, oversight, and accountability so that AI becomes a reliable part of the work rather than a disconnected experiment.

The ability to understand what AI is good for, where its limits lie, and how to critically evaluate the information it produces. AI literacy doesn't mean everyone has to know how to code. It means knowing how to use AI sensibly, safely, and purposefully. A good AI user doesn't just ask AI questions but also knows how to evaluate the answers: Is the information correct? What is it based on? Are perspectives missing? Can the answer be used as is?

An AI strategy isn't an IT project; it's a plan for how a company operates in a world where AI does part of the work. It answers these questions: where is AI given authority, where does a human decide, what data may AI use, which work steps are automated, and how is productivity measured? Without a strategy, AI use fragments into employees' random experiments — outside shared ground rules, information security, and metrics.

A good AI strategy starts with the business, not the technology. It defines how data, models, agents, integrations, and processes connect to one another and where in the work AI creates the most value. It also addresses what AI must not do, where human approval is required, and how costs, risks, and quality are kept under control.

An AI strategy isn't a list of tools. It's an operating model — and increasingly just as important as business strategy, digital strategy, or HR strategy.

The skill of phrasing instructions for AI so that the answers are as accurate and useful as possible. Good prompting is a fundamental skill for the future — like reading and writing in the digital age.

AI can produce information that sounds plausible but is completely wrong. This is one of the biggest risks in using AI. Always verify facts against a reliable source.

You shouldn't feed just any information into AI. You need to be especially careful with personal data, customer data, trade secrets, contracts, health information, and other confidential material. Data protection in AI use means the organization knows what information may be entered into AI, where the data is processed, whether it is stored by the service provider, and whether it can be used to train models. Without clear rules, a good tool can become a major risk.

An operating model in which a human reviews, approves, or steers what the AI does. AI can produce the first draft, analysis, or proposal, but a human is accountable for the final decision. This is especially important when AI is used in customer service, recruiting, healthcare, finance, law, or other situations where a mistake can affect people's rights or daily lives.

The AI Act, the EU's artificial intelligence regulation, is the European Union's regulatory framework for developing and using AI. Its goal is to ensure that AI is used safely, transparently, and with respect for people's fundamental rights. The AI Act classifies AI systems by risk level: some uses are prohibited, high-risk systems face stricter requirements, and many use cases come with transparency obligations. The regulation entered into force on August 1, 2024. The prohibited practices and the AI literacy obligation started applying on February 2, 2025, the obligations for general-purpose models on August 2, 2025, and most of the regulation becomes applicable on August 2, 2026.

AI Governance means the management, rules, responsibilities, and oversight of AI within an organization. It answers the question: who may use AI, for what purpose, with what data, with which tools, and who is accountable for the results? Good governance isn't just bureaucracy — it makes AI use safer, more effective, and more scalable. Without shared ground rules, every team may use AI differently, and then risks, costs, and quality can easily spiral out of control.

A technique in which AI first retrieves information from a database and then forms its answer based on it. It reduces hallucinations and keeps answers up to date.

Further training a pretrained AI model on your own data. This turns the model into a specialist in a particular domain.

AI doesn't read words; it breaks text into smaller pieces called tokens. Understanding tokens helps you grasp AI's limitations and costs.

The AI's "working memory" — how much text it can process at once. A larger context window makes it possible to analyze long documents.

Background instructions that define the AI's role and behavior in a conversation. The system prompt is a "hidden instruction" that guides every answer.

A technique in which the AI is asked to reason step by step before giving its final answer. It significantly improves the solving of complex problems.

An AI program that can act independently, use tools, and carry out multistep tasks. Agents are AI's next big step forward.

An autonomous agent is an AI system that can plan a task, use tools, and work step by step toward a goal without constant human guidance. For example, it can retrieve information, book calendar slots, write code, process documents, or send drafts for review. Skills are specialized functions available to the agent: one skill might be web search, another spreadsheet work, a third drafting emails, and a fourth updating a CRM system. The better the agent's skills, the more useful it can be — but the more oversight matters, too.

Orchestration means directing multiple AI models, tools, and agents so that they work together toward the same goal. One agent might retrieve information, another analyze it, a third write the report, and a fourth check the quality. An agent farm is a group of AI agents with their own roles and tasks — the idea resembles a small digital team: researcher, project manager, coder, and quality assurance specialist. Built well, this can speed up multistep work; built poorly, it can produce a lot of errors very quickly.

MCP, or Model Context Protocol, is an open standard that lets AI applications connect to external tools, data sources, and systems in a controlled way. You can think of it as AI's "USB-C connector": a shared interface through which AI can use documents, databases, calendars, project management tools, CRM systems, or developer tools, for example. MCP is especially important for agents and business use. When AI doesn't just answer questions but retrieves information, uses tools, and carries out tasks, you need a secure way to define which systems it can access and what it is allowed to do.

AI's ability to handle different types of data: text, images, audio, and video. Modern AI understands the world through many senses.

The value of AI doesn't come from trying it once out of curiosity. The greatest value comes when AI is connected to recurring everyday tasks: the ones that take time, require thought, or easily get left hanging.

1. Brainstorming and bouncing ideas around. Ask AI for options, perspectives, and questions. It works well as a starting point for your thinking, as long as you don't outsource the decision to it.
2. Drafting texts. Use AI for first versions of emails, social media posts, blog posts, reports, and presentations. A human finalizes the tone, facts, and point of view.
3. Summarizing. Let AI summarize long documents, memos, studies, or meeting notes. Ask separately for the key decisions, risks, and next steps.
4. Research and comparison. AI can help you map out options, gather background information, and compare solutions. It's still worth verifying the facts against reliable sources.
5. Supporting customer service. AI can suggest answers to frequently asked questions, summarize customer messages, and help keep a consistent tone.
6. Boosting sales and marketing. AI can help with sketching customer profiles, coming up with campaign ideas, personalizing messages, and polishing sales pitches.
7. Analyzing data. AI can help you spot trends, anomalies, and questions worth exploring next in your spreadsheets.
8. Writing code and finding bugs. AI can speed up programming, explain code, and suggest fixes. Final responsibility for functionality and security remains with the developer.
9. Learning and training. Ask AI to explain a difficult topic simply, create practice questions, or build a personalized learning path.
10. Automating processes. Connect AI to recurring workflows: reporting, document processing, classification, reminders, and internal information retrieval.

Lookup functions that let you find data in another table. XLOOKUP is the more modern and flexible version.

Excel's most powerful analysis tool. It condenses thousands of rows of data into summaries with just a few clicks.

Automate repetitive tasks by recording sequences of actions. VBA is Excel's programming language for more advanced automation.

Excel's data import and transformation tool. Import data from different sources, then clean and combine it automatically.

Automatic cell coloring based on values. It makes data visual and easy to read.

Presenting data visually with bar, line, pie, and other charts. The right chart type tells the story behind the data.