Security
Safe with your data before it’s useful.
How the AI Workspace keeps your information, your logins and your spending under your control.
Built in, not bolted on
It starts with no access, and earns it one door at a time.
The AI Workspace is built on open-source software that a large internet company made for its own staff, then released publicly in August 2026. It’s designed around one idea: an AI that’s useful at work has to be safe with the business’s data first.
Your team
Agent
Asks for what it needs
In between
Gatekeeper
- Read-only
- Only your jobs
- Bank details hidden
- Your OK before sending
Your business
Systems
- Job system
- Accounts
- Spreadsheets
Starts with no access
Agents and apps can’t see anything until you grant it, one system at a time.
Passwords stay out of reach
Logins for your tools live in the Gatekeeper. The agent and the code it writes never see them.
No wandering the internet
Code runs with outside internet switched off, except through the doors you choose to open.
Every read is recorded
If an agent has looked at payroll, it can’t then pass that on to someone who isn’t allowed to see payroll.
Spending limits built in
Every AI request is tied to a person or team, so you can set budgets and never get a surprise bill.
Not locked to one AI
Models are chosen through one gateway, so you can use the best one for each job and change later.
It’s yours, not ours.
Your workspace runs in your own account, with your data and your rules. The software is open source, so there’s no lock-in. If you ever stop working with us, it keeps running.
Book a free chat
Let’s find the first job to hand over.
No pressure, no jargon. Ask us anything in the chat: what it would cost, whether it fits your tools, or where we’d start. Then book a time to talk it through properly.
Custom Software Solutions · Cromwell, Central Otago