Agent07 security overview: data encrypted at rest with AES-256 and in transit with TLS 1.3 on Google Cloud Platform; OAuth-only access with scoped, instantly revocable tokens and zero stored passwords; organization-level isolation enforced with database row-level security and deal-scoped agent sessions; governed AI with human-approved learnings, holdout measurement, and trace capture; no training of foundation models on customer data; SOC 2 Type II certification in process; DPA and US/EU data residency options available for enterprise contracts; full data deletion on disconnection or cancellation.
Security & trustYour pipeline is the asset.
We built like it.
Agent07 reads your most sensitive commercial data — every email, call, and contract behind your revenue. This page is the plain-language account of how that data is protected, who can reach it, and what happens to it when you leave.
How it's protected
Four guarantees,
built in — not bolted on.
Encrypted end to end
All data is encrypted at rest (AES-256) and in transit (TLS 1.3). Keys are rotated automatically and never shared across tenants. Infrastructure runs on Google Cloud Platform.
OAuth-only access
Agent07 never stores passwords. OAuth tokens are scoped to the minimum required permissions and can be revoked instantly from your dashboard — access ends the moment you say so.
Isolation is architectural
Every piece of Deal Memory is walled to your organization at the database layer with row-level security, and agent sessions are scoped to a single deal. Nothing leaks across customers or teams — by construction, not by policy.
Governed AI
Agent decisions are traced and graded, learnings are human-approved before they apply, and their impact is measured against a holdout. We never train foundation models on your data — what Agent07 learns from your deals stays yours.
Compliance posture
Stated plainly,
including what's in progress.
Enterprise trust starts with not rounding up. Here is exactly where we stand.
Certification is underway. Current security documentation is available to enterprise customers on request.
A data processing agreement is available for any enterprise contract.
Regional data residency options for teams in the US and EU.
Disconnect a provider or cancel and your data is removed from our systems. We retain nothing after you leave.
The data lifecycle
What we touch, what we keep,
and what we delete.
What we access
Only what you connect, through scoped OAuth grants: email, calendar, meeting recordings, and the documents you attach to deals. No password ever touches our systems, and every grant is visible and revocable from your dashboard.
What we store
One Deal Memory per organization — emails, transcripts, documents, and the signals extracted from them — encrypted at rest and org-walled with row-level security. It exists to serve your agents, and only your agents.
What happens when you leave
Disconnect a provider and the data from that connection is deleted from our systems. Cancel your account and all your data is removed. We retain nothing after disconnection or cancellation.
Control & portability
It's your data,
and it's always in reach.
Your pipeline, your written knowledge and what the agents learn stay with your company — not the rep, not the AI model — in formats that open anywhere. Every action is on the record, and you can see exactly what an agent read before it wrote a word — no ticket, no waiting on us.
Export everything, whenever
Audit log → Export downloads your CRM as CSVs — deals, contacts, companies, activities — and your Brain as a folder tree of Markdown files. No ticket, no waiting on us, no export fee.
One log, and exports are in it
Audit → Record is a single filterable record of configuration changes, admin actions, field-level before-and-after edits, and every data export — including who ran it and what it contained.
Every agent run is on the record
Each run stores what context it read, which of your rules governed it, what it produced, and whether a human accepted, edited, or discarded the result. "Show me what the agent saw" is a question with an answer.
An agent can only read what you can
Permissions are applied at retrieval, not after it. A private note or a document outside your teamspace cannot surface in someone else's answer, or in a background job running on their behalf.
Security FAQ
The questions your CISO will ask.
Something we didn't cover? Write to security@agent07.ai and a human will answer.
Security your CISO will approve.
Setup your reps will love.
Connect Gmail or Outlook in 2 minutes — OAuth-only, revocable any time.