//legal · ai statement
AI Statement
Last updated 1 August 2026
This is the trust page. If you only read one legal document here, read this one.
1 · Agents assist over your data
The agents read the records you connect and prepare work from them — books, drafts, plans, reports. They act on your material, in your context, and explain what they did in plain language.
2 · Your data is isolated to your node
Everything you connect lives on your own node. It is not pooled, not mingled with another customer’s records, and not readable by anyone else’s agents.
3 · Shared agents never share data
The agents are shared and get better for everyone from how the stack is built and the material we author. They do not carry your numbers, customers or copy between accounts.
4 · Model-agnostic — bring your own AI
Connect Claude, ChatGPT or anything that speaks MCP. You keep your provider relationship, and you can change provider without losing your work.
5 · You control access (RBAC)
Access is granted per person and per area, checked on the server on every request, and fails closed. No grant means the data is never sent and the screen is never drawn.
6 · No training on your private data
We do not train models on your business data, and we do not sell it. Ever.
7 · Human oversight
Nothing posts, sends money or emails a customer without a person approving it. Where an agent is unsure, it stops and asks.
8 · Being wrong
Models make mistakes. Every number the agents produce is traceable to the source it came from, so you can check it rather than take it on faith.
Questions about any of this? hello@i2systems.ai