Zero Agents

What is Zero Agents?

Learn how roles, context, memory and collaboration shape an agent’s work, and how activity and durable agents contribute to tasks and wider processes.

Roles agents take on

  • Research

    Gathers and summarises what is known, from your data and from the sources you allow.

  • Doc Edit

    Drafts and edits documents, and brings in Research when it needs more to work with.

  • Chatbot

    Holds a conversation with a customer or a colleague, with the context of your business behind it.

  • Your own roles

    A support desk, a help desk, or a website assistant: a role you define, given your context and your data.

Business context
Data
Memory
Agent in a defined role
Completed task or flow step
Specialist agents collaborate on the work
An agent draws on business context to complete its role and contribute to a wider process.

Agents that make other agents better

  • Eval

    Scores an agent's work against what good looks like for that role.

  • Memory Extraction

    Keeps what is worth remembering from a piece of work, so the next run starts ahead.

  • Critic

    Reviews a draft and identifies improvements.

Doc Edit
Research
Research
An agent delegates research to specialists and brings their findings together. This is the supervisor pattern; delegation across several levels is the hierarchical pattern.

Activity Agent: short tasks

The default. It takes a task, finishes it, and hands back the result.

Durable Agent: long-running work

It stays with a piece of work over time. It can wait for a decision, and it can hand work to other agents.

These classes describe the duration and continuity of work.

What is an agent on Zero?

An AI agent with a defined role. It completes defined work independently or as a step in a flow, using relevant business context and data.

What roles can agents take on?

Research, drafting and editing documents, conversation, assessment, and first-line decisions. Zero provides roles such as Research, Doc Edit, and Chatbot, and you can define your own.

How does an agent know about my business?

Each agent is given business context, the data it needs through datasets, and memory of earlier work. Its output fits the way you operate because it starts from what your business already knows.

Can I define my own agents?

Yes. A support desk, a help desk, or a website assistant is a role you define, with your context and your data.

How do agents work together?

An agent can delegate focused tasks to specialist agents and combine their findings. In a flow, their work joins data, checks, and decisions in one process.

Agent

An AI agent with a defined role. It works on its own or as a step in a flow, with the context of the business and the data it needs.

Role

The job an agent is given: what it is for, what it may use, and what it hands on.

Context

What an agent knows about the business it is working for: its processes, its language, and the task in front of it.

Memory

What an agent keeps from earlier work, so the next run starts from what was learned.

Collaboration

Agents with defined roles share tasks and combine their work within a flow.

Zero Agents is part of Zero, the agentic orchestration platform.