Trust Center
How Quarza protects your data, your systems and your sign-off.
Security and compliance
Least-privilege access
Agents get only the access a task needs, through accounts you control and can revoke.
Your rules, respected
Built to work within India's DPDP Act and your sector's regulations.
Supervised by design
Every agent has a named human owner and an approval gate on irreversible steps.
Data governance and privacy
Your data stays yours
We don't use client data to train models.
Personal data masked
Personal data is masked or redacted before it reaches a model, wherever the task allows.
Leave with everything
Exit any time with your data, your workflows and full documentation.
Human-in-the-loop controls
Approval gates
Payments, customer messages and record changes wait for a person.
Off switch
Pause any agent instantly, from one place.
Daily reconciliation
Agent output is checked against your source systems every day.
The right model for every job
Best model per task
Claude, Gemini and GPT chosen for each job on accuracy, speed and cost.
Always current
We test new models as they ship and upgrade when they're better.
Failover
If a provider goes down, work moves to another model.
Documents
Shared under NDA with clients and prospects in evaluation.
- Security overview
- Data processing agreement
- Subprocessor list
- Incident response summary
Questions
Do you train AI models on our data?
No. We use model providers under terms that exclude client data from training.
Where is our data stored?
In your systems wherever possible. Where we host, we name the region in your contract.
Who can see what the agents do?
Your named owners and our delivery team. Every run is logged and reviewable.
What happens if an agent makes a mistake?
Irreversible steps wait for approval, outputs are reconciled daily, and we fix and document every incident.
Can we stop an agent immediately?
Yes. Every agent has an off switch you control.
