AI agents · The guide

AI agents: a plain guide to what they are and what they do

AI agents are software that can read, decide and act to finish a job for you. A chatbot only answers. An agent does the work: it fills the form, sends the email and updates the record. This guide shows what AI agents are, the main types, and how we build them into software.

How AI agents work

You give an agent a goal, not a prompt. It works out the steps, does them and checks the result. Every agent runs the same four moves.

It reads

The agent takes in what it needs: an email, a form, a file, a record in your system.

It decides

An AI model works out what the job needs next. This is the part a normal script can't do.

It acts

It uses your tools to do the step. It books the slot, writes the report or updates the record.

It checks

It looks at what happened, fixes what it can, and asks a person when it isn't sure.

The main types of AI agents

Single-job agents

One agent, one job, done well. Sort the tickets. Draft the reply. The easiest kind to test and trust.

Multi-step agents

One agent that runs a whole process with a few choices along the way, from first step to last.

Multi-agent systems

A team of agents that share one big job. One plans, one does the work, one checks it.

Chat agents

A chatbot that can also act. It answers the question, then does the thing the user asked for.

Voice agents

An agent your users talk to out loud, inside your app. It listens, answers and does the task.

Background agents

Agents that run on a timer or a trigger, with no one watching. They need the tightest limits.

AI agents we build into software

We don't sell agents off a shelf. We build them into your app or platform, so your users get the job done inside the product they already use.

Agents that get users started

A new user signs up. The agent sets up their account, pulls in their data and shows them the first win.

Agents that run the repeat steps

Fill, check, send, follow up. The steps your users do every day, done for them.

Agents that write the report

Your users' data, turned into the weekly report or client summary, ready to check and send.

Your program as an agent

Your method, built into an agent that coaches each client step by step, the way you would.

AI agents for Australian businesses

We're an Australian team. We've shipped 200+ apps since 2015, for founders, businesses and NSW Government. Here's what that means for your agent.

You own it all

The code, the AI accounts, the keys and the data sit in your name from day one. No lock-in.

Your data, planned first

We work out where your data sits and who can see it before we build, not after.

A person stays in charge

You decide what the agent does alone and what a person signs off. Every step is logged.

How to start with AI agents

  1. 01

    Pick one job

    Choose one repeat job that eats hours. Not the whole business. One job.

  2. 02

    Write down the steps

    List every step, every tool and every odd case. If you can't write it down, an agent can't do it.

  3. 03

    Build it small

    Build the agent for that one job and test it on real cases before anyone relies on it.

  4. 04

    Watch, then grow

    Read the logs. Fix what comes up. Then give the agent its next job.

AI agents: questions, answered

What are AI agents?

AI agents are software that can read, decide and act to finish a job. You give the agent a goal. It works out the steps, uses your tools to do them, and checks the result. A chatbot only talks. An agent does the work.

What is the difference between an AI agent and a chatbot?

A chatbot answers questions. An agent takes action. Ask a chatbot to move a booking and it tells you how. Ask an agent and it moves the booking, updates the record and sends the confirmation.

What can AI agents do for a business?

They take the repeat work off people. Common first jobs are sorting requests, drafting replies, filling records, writing reports and chasing the next step. The best first job is one you can write down step by step.

How much does it cost to build an AI agent?

A custom build is a fixed quote, and you see the running cost before you say yes. If you want a plan first, the Software Plan Workshop is $5,000, credited to a build. If you want a working first version fast, the 30-Day Prototype Launch is $18,400.

Are AI agents safe to use?

They are when you set limits. A good agent asks before anything that matters, hands over to a person when it's unsure, and logs every step. We test each agent on real cases before your users see it.

Do you build sales agents or voice agents for lead generation?

No. We build agents into software products. Sales agents and voice agents for lead generation are handled by a sister brand. Tell us what you need on a call and we'll point you to the right team.

Where do I start with AI agents?

Start with one job. Write down every step. If you want us to build it into your product, read our AI agent development page and book a product call. If you want to learn to build your own, read our how to build an AI agent guide.

Want an AI agent built into your product?

Book a free 30-minute product call. We'll work out which job an agent should take first, and what it would cost.