Platform / Security Security & Compliance

Security built for insurance agencies.

We never train or fine-tune a model on your data. AI drafts and proposes; a person confirms every change. Data changes and AI runs are recorded, and each agency's data is kept separate from every other agency's.

Preparing for a SOC 2 Type I audit.
Security Principles

Four commitments we build around.

AI Proposes. A Person Confirms.

AI drafts and proposes; a person confirms every change. Our agents can't change a record, send an email or contact a carrier on their own. Automations you set up run exactly as you configure them, and are logged.

Agency Data Separation

Each agency's data is logically separated, and every route is tested against cross-agency access. Your accounts, client profiles and placement history are never in reach of another agency.

A Record of What Happened

Data changes and AI runs are recorded, and the record can't be edited in ReliaPilot. Every AI run is recorded step by step. COI evidence packs and AI-activity exports give you documentation for your E&O carrier and compliance reviews.

We Don't Train Models on Your Data

We never train or fine-tune a model on your data. Models run inside AWS, not in public AI apps. AI runs on Amazon Bedrock inside our AWS account; model providers never receive your data. Your book of business is your asset.

Architecture

How we protect your data technically.

Encryption

Encrypted in transit. Integration credentials and sensitive intake answers are also encrypted by the application itself, and sensitive answers are masked, with every reveal recorded.

Logical Separation per Agency

Each agency's data is logically separated, and every route is tested against cross-agency access.

Role-Based Access Control

Admin and Director roles, plus per-user workflow access. Our support staff access your data only to help you, and their views of AI conversations are logged.

Recorded Activity

Data changes and AI runs are recorded, and the record can't be edited in ReliaPilot.

AI Inside Our AWS Account

AI runs on Amazon Bedrock inside our AWS account; model providers never receive your data. We store your records and AI history to run the service, and delete them on request.

Evidence for Your E&O Carrier

COI evidence packs and AI-activity exports, available when your E&O carrier or compliance team asks what happened and who decided.

AI Governance

Agents that prepare. People who decide.

Our agents work like new hires without a license: a defined job, limited access and a supervisor. Here's what that means in practice.

Agents look up, draft and propose

They can't change a record, send an email or contact a carrier on their own. A change an agent proposes happens only when a person clicks Confirm, under that person's permissions, and the confirmation is logged.

You can pause any agent, or all AI

Your admins can pause any agent, or all AI, for your agency, and stop a single run. Work in progress stops at its next step.

Every AI run recorded step by step

Which agent, which tools, which records it read, which model, and which version of your instructions it followed. Opening a run's prompts and responses needs audit permission, and each view is logged.

Certificate reviews your team confirms

COIVerify checks subcontractor certificates and endorsements against the contract's requirements and recommends a verdict. A person confirms it, and the review produces an evidence pack.

See how our agents work →

Common Questions

Questions we get from agency owners.

Can an AI agent send, submit or bind on its own?

No. Agents look things up, draft and propose. Sending, submitting and binding stay with your people.

Can we switch AI off?

Yes. Your admins can pause any agent, or all AI, for your agency, or stop a single run. Work in progress stops at its next step.

Who can read what the AI was asked and answered?

Only people with audit permission, and every view is logged.

Can my E&O carrier see what happened?

Yes. We can provide COI evidence packs and AI-activity exports. An evidence pack shows the documents received, the requirements version, the rulings in force, every check run, and every decision and who made it.

Where is my data stored?

In our AWS account. Each agency's data is logically separated, and every route is tested against cross-agency access. AI runs on Amazon Bedrock inside the same AWS account; model providers never receive your data.

Who can access my agency's data?

You and the team members you authorize. Our support staff access your data only to help you, and their views of AI conversations are logged.

Is my client data used to train AI models?

We never train or fine-tune a model on your data. Models run inside AWS, not in public AI apps. We store your records and AI history to run the service, and delete them on request.

What happens if a staff member leaves?

Access can be revoked immediately by your administrator. That user's prior activity stays on record. Memories they confirmed in Agency Brain stay with your agency — not with the individual.

Talk to our team

Questions about security? Talk to our team.

We'll walk you through how our agents work, what's recorded, and what a COI evidence pack contains — before you commit to anything.

Most pilots start within a week, and the first 30 days are free.