Typical cloud AI
- Your prompts and uploads are sent to a company's servers.
- Terms can allow your data to be retained or used for training.
- Hard to square with client privilege or privacy obligations.
- Ongoing per-seat or per-message subscription.
Private / local AI · the difference
AI that never leaves your building.
If you handle client files, health records, legal matters or financials, you shouldn't have to choose between using AI and protecting your clients. Private AI runs on your own hardware, in your own office, so you get the productivity without ever sending your data to someone else's cloud.
In plain English
What is private AI?
Private AI, also called on-premises or local AI, is artificial intelligence that runs on hardware you control: a computer or small server in your own office, instead of a third-party cloud service. Your documents and questions are processed on-site, so your data never leaves the building and is never used to train anyone else's model.
You get the same everyday usefulness people love about tools like ChatGPT: drafting, summarising, answering questions. But the information stays entirely in-house. For a business with confidentiality obligations, that's the difference between "we can't use AI" and "we use AI on everything, safely".
How it works
Your data stays put. The AI comes to it.
Runs on your hardware
An open, capable AI model is installed on a computer or small server in your office, not a cloud account, not a subscription that phones home.
Connected to your documents
I point it at your files, templates and policies so it answers with your knowledge: your letters, your matters, your way of doing things.
Nothing is transmitted
Prompts and files are processed locally. Nothing is sent away, stored elsewhere, or used for training. It can run fully offline if you want.
Your team just uses it
A simple, familiar chat-style tool. I train everyone so it's part of the day within a week, with no data-handling worries hanging over it.
The honest comparison
Typical cloud AI vs private AI.
Private / local AI
- Runs on your own hardwareNothing leaves the building.
- Never trains anyone's modelYour files stay your files.
- Fits confidentiality obligationsBuilt for privilege and privacy.
- Mostly a one-off setupRunning cost is closer to electricity than subscriptions.
Who it's for
Built for the businesses that can't just paste it into ChatGPT.
If a data leak would breach your professional obligations, or lose you a client's trust, private AI is how you get the productivity without the risk.
"We can't put client files in the cloud" is the most common sentence I hear. This is the answer to it.
Law firmsDraft letters & summarise matters, client privilege intact.
Accounting & financeWork with client financials that stay on-site.
Medical & allied healthPatient records never leave the practice.
Anyone sensitiveConsultants, advisers, government-adjacent. If it can't go to the cloud, it doesn't have to.
Questions people ask a chatbot
Private AI, answered plainly.
What is private AI?
Private AI (also called on-premises or local AI) is artificial intelligence that runs on hardware you control — a computer or small server in your own office — instead of a third-party cloud service. Your documents and questions are processed locally, so your data never leaves the building and is never used to train anyone else's model.
How is private AI different from ChatGPT?
With cloud tools like ChatGPT, whatever you type or upload is sent to a company's servers. With private AI, an open large language model runs on your own machine, so nothing is transmitted, stored elsewhere, or used for training. You get similar drafting and summarising ability, but the data stays entirely in-house.
Is private AI secure enough for client files and legal or medical records?
That is exactly what it is for. Because the AI runs on your own hardware and nothing is sent to the cloud, it suits businesses with confidentiality obligations — solicitors (client privilege), accountants and financial advisers (client financials), and medical or allied health practices (patient records under privacy and AHPRA obligations).
What can private AI actually do for a small business?
It can draft letters, emails and standard documents, summarise long files and threads, answer questions from your own documents and policies, and transcribe recordings — all using your business's own information, kept on-site.
What does it take to run private AI?
A reasonably capable computer or a small dedicated server in your office. I assess what you need, set up the hardware and models, connect it to your documents, and train your team. Ongoing running cost is mostly electricity, not a per-message cloud subscription.
Do you have to send my data anywhere to set it up?
No. The whole point is that your data stays with you. Setup happens on your own equipment, and once it's running the AI works fully offline if you want it to.
Want AI without handing over your clients' data?
Tell me what you'd want it to help with and how sensitive your data is. I'll come back the same business day with whether private AI is the right fit and what it would take.
Ask about private AI