Blog #1 AI Agents: What is an AI Agent?

Blog #1 | AI Agents explained: What is really behind them.

What is an AI agent? You hear everywhere that you need them and that AI agents are the future. But what exactly is an AI agent, and how do you decide if your company needs one? Over the past few months, we have had many conversations where agents were simply used to mean a chatbot with a better prompt. That is a shame, because that is not what they are. The technical difference is small, but the organizational one is huge. In our three-part blog series, we will take a closer look at AI agents. After reading this first part, you will know what components make up an agent, what the four collaboration models are, and which use case you can start with.

What is an AI Agent?

An AI agent is software that receives a goal and determines the path to achieve it independently. It breaks down the task, calls up tools, reviews the result, and corrects itself if something is not right. Only when the goal is reached or it gets stuck does it report back. The difference to everything we've seen before is summed up in one word: loop. A chatbot answers once. An agent works, checks its progress, and continues working. That is why it can complete a task, not just answer a question.

The stages can be described as follows:

Chatbot: The agent only knows what is in the model and what you enter. Useful for formulating, but blind to your company's needs.
Assistant: It sees your emails, files, and appointments. It formulates, summarizes, and suggests solutions. You have to execute them yourself.
Agent: It is authorized to act. It reads a ticket, searches the document repository, creates a response, and updates the system.
Agent Team: Several agents work together. One researches, one writes, and one reviews. They share interim results and escalate to you if anything is unclear.

Note: Most companies today are at Level 2 but believe they are at Level 3. If a person simply copies down every result or transfers it to the next system, that is Level 2.

What does an AI Agent consist of?

When we build agents, we always use the same six building blocks. This is true regardless of whether you are using Copilot Studio, Microsoft Foundry, or your own platform.

Building an AI Agent.

1. The Model

The language model provides the language understanding and planning capabilities. It is the most replaceable part of the system. Tying your agent design to a single model creates a problem for the next model update. We treat the model like a configuration, not an architectural decision.

2. Task and Instructions

Role, objective, boundaries, tone of voice, escalation rules. This is where most bad agents are created. A task like «Help with support» is not a task. «Answer questions about product X exclusively from the knowledge base, refer pricing questions to sales, escalate after two unsuccessful attempts» is a task.

3. Access and Tools

Without tools, an agent is just a conversation partner. With tools, they become a colleague. The standard that has become established for this is the Model Context Protocol (MCP). An MCP server makes a system – Microsoft Graph, Intune, a CRM, a time tracking system – usable for agents without requiring you to write a separate integration for every combination of model and system.

4. Memory

In the short term, the history of the conversation. In the long term, vector searches across documents and previous results. An agent without memory asks anew every day what your company actually does.

5. Identity

This is the part that endpoint and identity professionals understand immediately, and everyone else too late. An agent that takes action needs its own identity with its own permissions – not the account of the person who launched it. Microsoft introduced Entra Agent ID for this purpose, and Agent 365 uses this as its foundation. Without its own identity, you can't later determine who did what.

6. Guardrails

Approval steps for critical actions, logging of every tool call, rate limits, cost limits. Guardrails are not a hindrance, they are a prerequisite for allowing the agent more freedom.

How an AI Agent completes a Task.

The difference is best illustrated by a real-world example. Let's take a support ticket: «Outlook will not start after the update, and I cannot access my emails». Here is how an agent would handle it:

Understand: They read the ticket, identify the device name and user, and note that information is missing: Which update? Which Outlook version?
Plan: They establish a sequence of steps: Check the device in endpoint management, review the update history, and search the knowledge base for known issues.
Use tools: They query the device, retrieve the installed updates, and simultaneously search the documentation for the same error pattern.
Check: The search returns three results, two of which do not match the version. So they discard them. This is the part a pure chatbot cannot do.
Act or escalate: If a known issue matches, they add the solution to the ticket and mark it as «waiting for feedback». If none match, they escalate the issue with a summary of what they have already checked.

The value lies in steps 4 and 5. An assistant would have stopped in step 3 and shown you three links. An agent assesses and decides whether they can continue working. And step 5 contains the most important design decision: What can they do themselves, and when does the task need to be delegated to a human?

Note: When escalating a task, make sure that an agent always provides information about what they have already reviewed. Otherwise, the human will have to start over, and the time saved will be lost. This is a matter of how the task is worded, not a question of the model itself.

Skills: How an Agent acquires Skills.

Tools give an agent access. But access is not the same as ability. An agent with access to Word does not know what your offer looks like. Skills fill precisely this gap. And they are the most underrated innovation of the last year. A skill is, at first glance, unremarkable: a folder containing a file called SKILL.md. This file describes, in plain language, how to perform a specific task. Optionally, it also includes scripts, templates, and reference documents. Anthropic introduced the concept in October 2025 and released it as an open standard in December 2025. According to the official directory, around forty products now support the same format – including OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, and VS Code. So, a skill you write once will work in multiple tools.

The real key is how they are loaded.

Discover: At startup, the agent only reads the name and description of each skill. This costs approximately 80 tokens per skill.
Activate: If the skill matches the task, it reads the complete SKILL.md file. This is where it gets expensive.
Execute: It follows the instructions and uses the provided scripts and templates.

The effect is enormous: An agent can know dozens of skills while consuming less context than a single, fully loaded skill. This solves a problem that was previously almost impossible to solve: providing an agent with a wealth of company-specific knowledge without overloading its context. For companies, the practical value is even greater than the technical one. A skill is the first artifact form in which you can document process knowledge in a way that both humans and agents can use it. It is version-controlled, reviewable, and has an owner. If you start an agent project today, ask yourself early on: Which five skills do we need, and who in the business unit will write them?

Note: Write skills like a good set of instructions for a new colleague. When does it apply, what are the steps in which order, and what is prohibited? The description in the file header is the most important part. It is what the agent uses to decide whether to even read the skill.

How an AI Agent acts: 4 Ways.

«Tools» is a collective term, and the differences between them are crucial. An agent can act in four different ways, and in a well-designed system, it combines them.

1. Integrations

The clean approach: The agent calls an interface – Microsoft Graph, CRM, time tracking, ticketing system. Permissions are clearly defined, every call is logged, and the result is structured. This is now standardized via the Model Context Protocol: A system is made usable for agents once and then used by every agent. This is the crucial part for operations. The toolkit becomes a manageable list that can be enabled or disabled. Where an integration exists, it is always the first choice.

2. Command Line

Not every system has a usable interface, but almost every system has a CLI. An agent authorized to use the command line can execute PowerShell scripts, analyze logs, generate configurations, and run tests. This is both the most powerful and the most dangerous approach: a single command can do a lot of damage. Therefore, an allowlist of permitted commands and a sandbox environment are essential. More on this in Part 3 of our blog series on AI agents.

3. Browser and Computer Use

The solution for everything that has neither an interface nor a CLI: The agent is given a browser or an entire desktop session, sees the interface as an image, and clicks and types itself. This removes the last technical barrier, because essentially an agent can take over everything a user can do on a computer and in a browser. Researching, filling out forms, using portals, transferring data from one application to another, creating and saving files. And in many of these tasks, it is not only just as good, but faster and more thorough, because it does not get tired and does not skip any steps. The point that is usually missing in the discussion: A human works on one thing at a time. Agents work on many things simultaneously. Ten portal operations, twenty research projects, fifty documents – this all runs in parallel, not sequentially. This is precisely where the difference lies, which is reflected in the numbers: not in the fact that one task takes five minutes longer, but in the fact that one person can complete a hundred tasks in that time. In practice, this is the solution for legacy systems, supplier portals, and government applications – anything where your process currently relies on someone manually typing something in. However, it is also true that this is the most vulnerable of the four methods, then a changed layout can break a workflow. Therefore, implement validation steps instead of blindly trusting.

4. Skills

The other three paths are access. Skills are abilities. Skills tell the agent how to use the access points within your organization effectively. An agent with browser access to a supplier portal is useful. An agent who also has a skill for «how we order» is fully operational.

What an AI Agent is not.

Let's be clear about this to avoid a lot of disappointment:

An AI agent is not a replacement for a sound process. If your process is broken, you are just automating the chaos.
An AI agent is the wrong choice for tasks with fixed rules. Approval based on four fixed criteria belongs in automation, not in a language model.
An AI agent is not a replacement for good data. It can only find what is discoverable and has legitimate access.
An AI agent is not a staff reduction program. The cases that work save time – they do not replace roles.

Why your Company needs AI Agents.

So far, it is been about technology. Now for the part that drives the decision. For its Work Trends Index 2026, Microsoft surveyed around 20,000 knowledge workers in ten countries and analyzed anonymized usage signals from Microsoft 365. Three figures are truly relevant:

The number of active agents has grown 15-fold year-over-year – and even 18-fold in large companies.
46% more executives say their organization uses AI agents to automate entire workflows or business processes.
71% more employees in so-called frontier firms say their company is successful, compared to 37% globally.

The figure that impressed us most is a different one: According to the report, organizational factors explain around 67% of the impact of AI, while individual factors account for only about 32%. Put simply: It is not how skillfully individual people prompt that determines the outcome. It is whether the company transforms work. And at the same time, only 26% of executives say their organization has a clearly defined AI strategy. That is where the competitive edge lies.

Note: The term frontier firm comes from the Work Trends Index. It refers to a company that purchases intelligence on demand, works in human-agent teams, and establishes a new role: the Agent Boss. A person who builds, commissions, and manages agents.

The four Patterns of Collaboration.

The report of 2026 describes four patterns of how people and agents collaborate.

The leap from Editor to Director is the real step. It is not a technical issue, but a question of trust and governance: You are delegating an entire task, no longer just a draft. This requires protocols, permissions, and a defined escalation path; otherwise, delegation becomes a loss of control.

Which Use Case should I start with?

Three criteria: Does it happen often? Is the data located in a place I have authorized access to? And is an error noticeable and correctable before it causes problems? If the answer to any of these is no, it is not a good first use case.

How do I calculate the benefits?

«Saves time» does not survive any budget round. Therefore, calculate before building, and be deliberately conservative. Example: Ticket triage in a service desk with 60 tickets per day: Today, pre-sorting and enriching a ticket takes about four minutes. That is four hours per day. Realistically, the agent pre-processes 60% of the tickets completely, makes a mistake on 15%, and a human corrects it. That leaves about two hours per day – not four. The corrections take time and should be factored in. Consider this in terms of model and platform costs, as well as the one-time setup. If it does not pay for itself in under six months, it is the wrong first use case. This calculation is not pretty, but it is defensible. And it almost always leads to a narrower definition of the use case. Which is good, because narrower agents perform better.

What you need to clarify beforehand.

Inventory: Which agents are already running in your tenant? The answer is almost always «more than you thought.»
Remember: Without a registry, there is no governance.
Identity: Does each agent have its own identity with its own set of permissions? If not, change that before going live.
Data classes: Which data is an agent allowed to see, and which is strictly off-limits? This is a decision for the business department, not IT. Measurement: How will you know in eight weeks that it is working? If you cannot say now, you will not be able to later.

Mistakes you can avoid.

Start with the most intriguing case instead of the most useful one. The most intriguing case is usually the one where everything is missing: data, permissions, process.
Build without a way back. Every action needs a way to undo it – reopening a ticket, saving an email as a draft instead of sent. Without a way back, every mistake becomes an incident.
Start with a messy knowledge base. If two documents contradict each other, the agent will contradict itself too. And you will look for the error in the model, even though it is in the folder.
Do not set a budget limit. Agents run in the background, and consumption models don't remember that you wanted to save.

Conclusion.

An AI agent is not simply a better chatbot, but an actor with a mission, tools, and an identity. Whether your company needs one is not determined by the technology, but by whether there is recurring work, data that is readily available, and where you can identify errors early. If so, start with precisely such a case, give the agent its own identity, and measure the results.

FAQ: Frequently asked Questions about AI Agents.

A chatbot answers a question once. An agent receives a goal, plans the route itself, calls up tools, checks the result, and continues working until the task is completed or it gets stuck. The difference lies in this checking and correction step.

Four things:

an inventory of which agents are running,
a separate identity with its own rights for each agent,
a decision by the department regarding which data an agent is allowed to see,
and a benchmark after eight weeks to determine if it worked.

With one that occurs frequently, whose data is stored in an accessible location, and where an error is noticeable and correctable. In practice, this usually involves ticket triage, preparing quotes, or research and summarization.

The issue is not so much the license as the resource consumption: An agent running hourly in the background costs significantly more than one that operates on a click-by-click basis. Calculate conservatively for correction costs and set a budget limit for your total costs (but not for tokens per session) before the first agent goes live.

Jannik Reinhard, Head of AI at Epic Fusion

Want to know when AI Agents would make sense for your company?

I would be happy to help you assess this. Feel free to contact me for an initial consultation, and we can look at specific cases from your daily work.

Contact me

What is an AI Agent? And why does a Company need AI Agents?