Command 
CanTicketOur AI agents run in Australia.
Command Centre is in pilot. The agents we are training with our pilot businesses run on Amazon Bedrock in Sydney and Melbourne, so every draft, plan and answer they produce is processed in Australia. Pilot data and its backups are stored in Sydney, and the software blocks any AI request that would leave the country.
Two ways to power your agents
Configured by us
In the pilot, we set up and train your agents on our own Amazon Bedrock account in Australia. Nothing to configure: the AI runs in Sydney and Melbourne from day one.
Bring your own key
Prefer your own AI account? Connect it and pay your provider directly; we never resell AI. Amazon Bedrock in Australia is the one we recommend, and we can set it up with you. If you choose a provider overseas, your AI runs where that provider runs.
What stays in Australia
The agents’ thinking
Every draft, plan, briefing and answer our pilot agents produce is processed on Amazon Bedrock in Sydney and Melbourne.
Your records
Your company brain, conversations, plans and reports are stored in Sydney, each business in its own separate database.
Your backups
Copied every second to storage held in Sydney, separate for each instance.
What still leaves, and why
We would rather tell you than have you find out. None of these is where your business’s conversations or the agents’ work are kept.
Sign-in
Your name and email address, held by our sign-in provider in the United States.
Encrypted disk copies
Daily copies of the server disk, held by our hosting provider in the United States, until our Sydney copy replaces them.
Technical logs
Our software’s running diary: page requests, errors and scheduled jobs. We write them; our hosting provider stores them and does not say where. They can include an email or IP address, but they are not where conversations or the agents’ work are kept. We are moving them to storage we choose in Sydney.
Sends you approve
An email, post, invoice or update you approve goes out through your own accounts, such as Gmail or Google Workspace, Microsoft 365 and Outlook, Facebook and Instagram, Xero, and your CRM. Each keeps it wherever that provider normally does.
How we keep it that way
Enforced in the software
Command Centre only sends AI requests through Amazon’s Australian route. A request that could be routed overseas is refused before anything is sent. If the setting is ever wrong, the AI switches off rather than going overseas.
Locked at Amazon too
The access key Command Centre uses is limited by Amazon to the Australian route and the two models we use, so the rule holds even outside our software.
Not used for training
Amazon states Bedrock does not use prompts or answers to train models, and the model makers cannot see them.
People approve what goes out
Agents draft; a person approves every email, post and update before it is sent.
Governance you can ask to see
We run Command Centre the way we ask our clients to run AI: an AI advisory board that advises and a person who decides, written policies adopted with signatures, an AI register, an incident process, and regular audits. We align our practice with the Australian Government’s Guidance for AI Adoption (AI6), action by action, and keep the evidence. Command Centre is listed on the National AI Centre’s AI Directory.
See it run live →Questions
Does my data leave Australia?
Pilot records and their backups are stored in Sydney, and the work our pilot agents do is processed on Amazon Bedrock in Sydney and Melbourne. A few supporting services sit overseas: sign-in names and email addresses, encrypted daily copies of the server disk (until our Sydney copy replaces them), and our software's technical logs, which can include an email or IP address. When you approve a send, it goes out through your own accounts, such as Google, Microsoft 365, Facebook and Instagram, Xero or your CRM, wherever that provider keeps it.
Is my data used to train AI models?
No. Amazon states that Bedrock does not use prompts or answers to train models and that the model makers cannot see them. Training our agents means teaching them your business through their brain and your feedback, not training an AI model on your data.
Which AI models do your agents use?
Anthropic's Claude, run by Amazon Bedrock in Australia. A larger model handles leadership work such as the managing director and the advisory board; a smaller, faster one handles everyday agent tasks. Matching the model to the job keeps costs down.
How do you stop AI requests going overseas?
In the software itself. Command Centre only sends AI requests through Amazon's Australian route (Sydney and Melbourne). A request that could be routed outside Australia is refused before anything is sent, and if the setting is ever wrong the AI switches off rather than going overseas. The access key is also limited by Amazon to the Australian route.
What about businesses that run their own Command Centre?
They bring their own AI account and pay their provider directly; we do not resell AI. We recommend Amazon Bedrock in Australia and can set it up with you.
Are you AI6 compliant?
AI6 is the Australian Government's voluntary guidance for AI adoption, so there is no compliance certificate. We align our practice to it action by action, keep the evidence, and will publish our progress as our policies are adopted. We never claim to be compliant or certified.