From model to agent
You see what an agent adds to a model on its own: tools, a loop of actions and agents working in parallel.
ObjectiveBy the end of this lesson, you will be able to describe an agent’s loop and read, in the steps feed, the tools and parallel agents its result rests on.
A model on its own only writes. An agent is a model placed in a loop: it reads your request, decides on an action, carries it out with a tool, reads the result, then decides what comes next. The loop continues until the task is done, or until it needs you.
Tools change everything
- Web and academic search brings up-to-date information, with its sources.
- Reading your files and your knowledge library brings your context.
- The controlled browser can consult pages and fill in forms, under your control.
- Spreadsheets and code make it possible to calculate correctly.
- More than forty skills bring specific know-how: analysis, writing, compliance, audio transcription, image generation.
Learnya shows its plan and every tool it uses. You therefore see, step by step, where the result comes from. This is a major difference from a model on its own, whose answer comes out in one piece without you knowing what it rests on.
Several agents in parallel
For a complex task, Learnya can delegate to several agents working at the same time. One explores a topic, another checks the sources, a third looks for what contradicts the conclusion. Their results are then brought together. This set-up mirrors a small team that reviews its own work.
What does not change
An agent is still a language model. It can misunderstand a request, choose the wrong tool or draw a hasty conclusion from a web page. That is why Learnya stops before any binding action, such as sending, buying, publishing or deleting, and asks for your approval.