Now with voice โ€” Introducing the AI Voice Agent
Security & Privacy

Security and privacy,
built in from the start.

You're trusting AI Employee with your documents and, for authenticated self-service, your members' account data. Answers stay grounded in your own content, data is encrypted and access-controlled, every interaction is logged, and you stay in control of what's shared and retained.

Our commitments

The principles we
design around.

These aren't add-ons โ€” they shape how the product is built and operated.

๐Ÿ“‘

Grounded, not guessing

Answers come only from your own documents and data, with the source cited. When the answer isn't there, it hands off to your team instead of inventing one.

๐Ÿ”’

Your data stays yours

Your documents and member data are used to serve your deployment โ€” not sold, and not used to train public, shared models.

๐Ÿ—๏ธ

Least-privilege access

Account data is reachable only after a member authenticates, and only the data needed to answer the question is accessed.

๐Ÿงพ

Everything is logged

Every question, answer, source, and action is recorded โ€” so a disputed interaction has a complete, reviewable record.

๐Ÿ™‹

A human stays in control

Sensitive or irreversible steps pause for staff approval, and your team can review, correct, and configure what the assistant does.

๐Ÿ›ก๏ธ

Privacy by design

Consent, data-access controls, and retention settings are part of the product โ€” built with CCPA-style requirements in mind.

How your data is handled

Protected in transit,
at rest, and in use.

Practical safeguards across the lifecycle of your data โ€” from the moment a document is uploaded to the moment an answer is given.

โœ“ Encrypted in transit (TLS) and at rest
โœ“ Hosted on reputable, access-controlled cloud infrastructure
โœ“ Role-based access and least privilege internally
โœ“ Member account data accessed only after secure authentication
โœ“ Each community or location kept logically separated
โœ“ Configurable data retention and deletion on request
Where your data lives

Hosted in the
western United States.

Where each part of the service runs, and how staff sign in to it.

WhatWhere
Application and databaseGoogle Cloud, Oregon (us-west1): Cloud Run and Cloud SQL for PostgreSQL
Uploaded documentsGoogle Cloud Storage, Oregon (us-west1)
Search indexQdrant Cloud on AWS, Northern California (us-west-1)
Website and consolesCloudflare
โœ“ Data at rest encrypted by Google Cloud with Google-managed keys
โœ“ Database backed up daily, with 7 days of backups kept
โœ“ Staff passwords stored as bcrypt hashes, never in plain text
โœ“ Staff sessions expire after 30 minutes and renew for up to 7 days
โœ“ Optional staff sign-in with Google
โœ“ Support access to your account is read-only, time-limited and logged
Service providers

Who processes
your data.

These providers process customer data on our behalf.

ProviderWhat it doesWhat it receives
Google CloudHosting, database and file storageAll customer data
OpenAIWrites answers and builds the search indexQuestions and the document passages relevant to them
Qdrant CloudHosts the search indexDocument passages
Google (Gemini)Voice conversations, when voice is turned onSpoken questions and relevant document passages
AnamVideo avatar, when live video is turned onThe conversation during a video session
ResendSign-in, invitation and verification emailsStaff and board members' email addresses
CloudflareHosts the website and consolesVisits to those sites
Privacy & your obligations

Built to support your
privacy commitments.

Homeowners, families, and members have rights over their data, and you're responsible for honoring them. AI Employee is designed to make that straightforward.

โœ…

Consent built in

Disclosure and consent can be presented before any personal data is collected or an authenticated action is taken.

๐Ÿ“จ

Access & deletion requests

We support fulfilling data-subject requests โ€” including access and deletion โ€” for the data held in your deployment.

๐Ÿงญ

You control the scope

You decide which documents, systems, and data the assistant can use, and can change or revoke that at any time.

Common questions

Straight answers
on security.

Do you use our data to train AI models?

No. Your documents and member data are used to power your own deployment. They are not sold, and not used to train public, shared models.

Can the assistant make up an answer?

It answers only from the content you provide and cites the source. If the answer isn't in your documents, it says so and routes the question to your team rather than guessing.

How is members' account data protected?

Account-specific information (balances, payment history, registrations) is only accessible after the member securely authenticates, and only the data needed to answer is retrieved.

What happens to the record of a conversation?

Every interaction โ€” question, answer, cited source, and any action โ€” is logged so it can be reviewed later, with retention you can configure.

Where does it run relative to our existing system?

It layers on top of the platform you already use, which remains your system of record. We don't replace it or take over your data.

A note on certifications. We follow the security and privacy practices described here as we work with our first design partners. If your organization requires specific documentation, a data-processing agreement, or a security questionnaire, reach out and we'll walk through it with you directly.

Have a security or
privacy question?

We're happy to walk through how AI Employee handles your data โ€” and to complete a security review with your team.