Your data stays yours.
Where the system runs, who can see what, and what happens to your recordings and records. In plain language, before you sign anything.
Quick answer
With Trilix AI, your business owns its data: call recordings, camera data, reports and customer records. We use them only to run your system. Hosting can be managed by us, kept in Canada, the US or Europe, or run in your own cloud account. Access is set by role, actions are logged and sensitive steps wait for a person.
Where it runs
Most clients use our managed hosting. If your business needs more control, the same build can run closer to home.
Managed cloud
The default. We host and run it for you, so there’s nothing for your team to install or patch.
A region you choose
Hosted in Canada, the US or Europe when your customers or contracts need the data kept there.
Your own cloud account
The system runs inside an account your business owns, and you hold the keys.
Your own servers
For stricter environments, parts of the system can run on hardware at your site.
How we handle it
You own the data
Call recordings, camera data, reports and customer records belong to your business. We use them only to run your system.
People see what their role needs
Owners, managers and staff get different views. A store manager sees their store, not the group’s payroll.
A record of what happened
Calls, bookings, alerts and automated changes are logged, so any action can be traced back.
People take the sensitive cases
Anything outside the rules you approved goes to a person, with the context attached.
Scripts you sign off on
What an agent says and does is written down and approved by you before launch.
Privacy for employees
Where conversations are recorded for coaching, employees can see their own recordings and managers see graded reports.
Connected the safe way
We use each system’s official connections or exports, with access limited to what the build needs.
Watched after launch
We monitor the system once it’s live and fix problems as part of the monthly fee.
What safeguards are built in?
Recordings, consent and AI models are where most of the questions come from. This is how we handle each one.
A disclosure on every call
Before an agent answers its first call, we agree the opening line with you: that the call may be recorded and that the caller is speaking with an automated agent. Callers who ask for a person get transferred.
Outbound calls only with consent
In 2024 the FCC ruled that AI-generated voices count as artificial voices under the TCPA. We only set up outbound calls to people who have agreed to hear from you, and we keep a record of that consent.
Recording rules where you operate
Several US states require consent from everyone on a recorded call, and Canadian privacy law expects notice and consent. Staff and customer notices are set up with you before anything is recorded.
Employee recordings stay private
Where conversations are graded for coaching, the raw recording can be limited to the employee in it, while managers see graded reports. For one Texas client we also drafted the employee agreements around Texas law.
Read-only where that’s enough
If the system only needs to read from a tool, the connection is read-only. The assistant we built for a UK holding group reads Jira and Slack and changes nothing in either.
A second AI checks the first
Where it matters, an audit agent reviews the main agent’s work on a schedule. At a multi-store quick lube, that audit caught the first model ignoring some booking rules, and we moved to a stronger model.
Answers with sources
Knowledge assistants answer from your approved documents and show where each answer came from. If there’s no verified answer, they say so.
Model providers in writing
Your data is used only to run your system. Which model providers are used, and under what terms, is written into the scope so your team can review it before anything is built.
Data and security questions
Something else on your mind? Ask us directly or call 416-262-2723.
Who owns the data a Trilix system collects?
Your business does. Call recordings, transcripts, camera data, reports and customer records belong to you. We use them only to run your system.
Will our data be used to train AI models?
Your data is used only to run your system. Which model providers are used, and under what terms, is part of the written scope and the security review, so your team can check it before anything is built.
Where is the data hosted?
By default on our managed hosting. It can also be hosted in Canada, the US or Europe, inside a cloud account your business owns, or partly on hardware at your site.
Who can see the recordings and reports?
Access is set by role. Owners see the group, a store manager sees their store, and employees can be the only ones able to play back their own recordings.
Can the AI take actions without a person approving them?
Only the ones you approve in the spec. Postings to the books, refunds, messages to customers and anything outside the approved rules can wait for a named person, and every action is logged.
Do you hold security certifications?
If your procurement team needs a specific certification or contract term, ask on the first call and we’ll tell you plainly where we stand. Send any security questionnaire early. We’d rather answer it before scoping than after.
Keep reading
- Enterprise AIGovernance, hosting choices and how procurement works with us.
- AI receptionistRecording, consent and disclosure on every call.
- Custom AI voice agentsInbound and outbound agents, and the consent rules for outbound calls.
- IntegrationsThe systems our client builds connect to today, and how.
- AI glossaryPlain definitions of data residency, guardrails, PIPEDA, GDPR and more.
- Privacy policyHow this website handles your information.
Bring your security questions to the first call.
Tell us what your customers, insurer or IT team need to see. We’ll walk through where the data goes and who can access it.