Skip to content
Tekoälli

Layer 4

Applications

How a model becomes something that is actually of use to somebody.

On this page

A language model on its own is not a product. The application layer is where the benefit is decided: who may see what, where the correct information is found, when a person has to approve the result, and who answers for a mistake.

What happens in this layer

  1. 01

    The prompt says what is being done right now

    Instructions and context: the user's question, the application's rules, examples, documents, and a description of the tools available.

  2. 02

    Retrieval brings in the facts

    Before answering, the system searches the company's own documents and hands the passages it found to the model. The model itself does not change, but the sources can be shown to the reader.

  3. 03

    Fine-tuning changes how it behaves

    The model's numbers are adjusted for a purpose: a house style, a classification task, the conventions of a particular field. A poor way to add fresh facts.

  4. 04

    An agent is a loop

    The model sizes up the task, picks a tool, looks at the result, corrects course and continues. The tools, the permissions and the checks decide the quality, not the eloquence.

A common misconception

RAG
The model looks the facts up before answering, so the answer can show its source.
Fine-tuning
Changing a model's weights for a particular purpose.
Prompt
The instructions and context given to a model.
Agent
A program that does not merely answer but acts: picks a tool, looks at the result and corrects course.
Guardrails
The rules and checks that limit dangerous or incorrect actions.
Hallucination
The model produces a convincing but false statement.

Articles on this topic

Vibe coding and AI-assisted programming

Two different ways of working get grouped under one label, though the distinction that matters is whether the generated code is reviewed. What each approach is good for, what remains after the demo, and what productivity research actually shows.

20 min

AI use cases in plain language

Thirty concrete use cases, from the home to the forest, from the production line to a sports club. What the machine does in each, where it goes quietly wrong, and what happens to a use case after it is chosen.

18 min

Agents, orchestration and use cases

What a multi-agent system actually is, the four structures you find in one, what each is good for, and when a single agent is enough.

19 min

What an agent actually is

An agent is not magic. It is a loop: the machine tries something, looks at the result and corrects.

3 min