Do I need to know how to code to create an agent?

Describe the role in one sentence and Skyller writes the name, personality, instructions and opening messages for you to review; or fill in five steps by hand. Choosing documents and switching on tools is a screen of choices, with no code.

Is an agent an AI model?

The agent is the configuration: instructions, documents, tools, AI level and reach. The model, from providers such as OpenAI and Anthropic, is the engine that writes the answer, and Skyller chooses which one to use for the level and the task.

What is the difference between an agent and a team?

An agent is a specialist with its own instructions and tools. A team brings several agents under a leader, with a way of working: coordinate, send to just one, send to all or split into tasks. In the conversation, you choose whether to talk to an agent or a team.

Is there a ready-made agent for my area?

Skyller keeps a catalogue of agents by area — accounting, operations, customer service, sales, HR, logistics, legal, marketing, IT and others — that appears next to yours. Use one as it is or duplicate it to adapt; what counts is what appears in your catalogue.

Platform · Agents and teams

Whoever knows the routine builds the agent. The team uses it.

The agent carries the instructions, the published documents and only the tools the company released. Teams bring several agents to one goal.

Talk to a specialist Try it free

See how it works

Skyller · Create agent · review configuration

Create agent · review configuration

Onboarding assistant

Goal

Audience

How the agent should behave

Create an outline to {{objetivo}}. Consider the needs of {{publico}}.

Use the hub’s published documents and cite the source. Ask for review when information is missing.

Demonstration · illustrative data

AUTOMATIC · ADVANCED · PREMIUM

the agent carries the level; Skyller chooses the model

PERSONAL · COMPANY · PLATFORM

your agents, the company's and Skyller's ready-made ones

TOOL BY TOOL

each agent uses only what was switched on for it

4 TEAM MODES

coordinate, route, send to all and tasks

Where this shows up in your day

The knowledge of the routine lives in one person's head.

Finance

The same question, answered three ways

Each person asks the chat in their own way and gets a generic answer, without the credit policy or the order system.

One agent for the area

HR

Explaining the policy for the tenth time

Holidays, remote work, onboarding: the questions repeat, and the answer changes with whoever was asked.

Answers by the policy

Operations

Work that passes through several hands

Gathering data, checking the procedure, drafting the summary: each step calls for a different specialist.

One team, several steps

How it works

From what one person knows to an agent the team uses.

01
Describe or build
Describe the role in one sentence and Skyller writes the name, personality and instructions for you to review. Or build it by hand, in five steps.
02
Give it knowledge
Choose the documents it always checks and the hub where it works. Drafts and expired documents stay out.
03
Switch on the tools
From the connected systems, switch on only the functions it needs and mark which ones ask for approval before acting.
04
Choose the AI level
Automatic, Advanced or Premium. The agent carries the level; Skyller chooses the model for each task.
05
Share or publish
It starts as yours alone. Invite people or groups as readers or editors; each one uses it with their own access.

Live proof

One request, three agents, one answer with its source.

Starting point
The “Monthly close” team, in Coordinate mode, brings together three agents: one checks the Finance hub documents, one checks the closing procedure, one drafts the summary.
What happens
1. 01In the Finance hub chat, the person picks the team and asks: “Prepare a summary of September's outstanding items.”
2. 02The leader splits the request among the members, and the conversation panel shows each one at work.
3. 03The leader combines everything into a single summary, with the procedure cited as a link.
Result
An answer built by specialists, in the same chat, and whoever asked sees what each one did.

Skyller · Team · Monthly close

Prepare a summary of September’s outstanding items.

…

Document analystRead published procedure…

Financial reviewerReview outstanding items…

Summary writerConsolidate the result…

Skyller is an AI and can make mistakes. Remember to check.

Responding

Demonstration · illustrative data

Good to know

The team works with what each member can access and flags what was missing.
A sensitive action in a connected system waits for approval, including when a team is the one asking.
Each run starts from a request or from an automation with an owner.

Fictional company, agents and documents; the team modes and the panel are the product's.

To evaluate calmly

What an agent carries, and who is in charge of it.

Select a topic to see its capabilities, conditions, and limits.

Instructions, personality and opening messages: Name, avatar, description, category, instructions and conversation starters, all editable later. Creating with AI costs one normal turn at the Automatic level.
Fixed knowledge base: The published documents it always checks, plus what it finds in the hub and workspace of the conversation.
Tools switched on one by one: Within what the company connected, each item switches on or off and has an approval rule: inherit, always ask or never ask. The requirement can come from the company, the agent or the conversation.
AI level and reasoning: Automatic lets Skyller choose the model; Advanced and Premium fix the tier. Reasoning follows the level.

Real application

The finance agent, built by the finance person.

01Finance analyst
Describes the agent in one sentence: answer questions on credit and collections and look up customer history. Reviews what Skyller proposes and saves.
02Finance analyst
Picks the published Credit and Collections Policy as the fixed base and places the agent in the Finance hub.
03Administrator
In the CRM already connected, switches on only “look up customer history” and “send email to the customer”, and marks the send to always ask for approval.
04Finance analyst
Shares it with the Finance group as readers and with the coordinator as editor.
05Finance team
Asks in the hub chat: the lookup comes back at once; sending an email to a customer opens the approval card before it happens.

Questions from whoever decides

What people ask before building the first agent.

Next step

Build the agent for the routine that drains your area most.

Describe the role in one sentence, choose the document that counts and switch on only the functions needed.