Put autonomous defense on your own estate for fourteen days.
Not a sandbox and not a slide deck. The production platform, your assets, your identity provider, and a written report at the end that says what it found and how fast it contained it.
Instant agent deployment • Zero friction integration • No production change window
How the fortnight runs
- Day 0
Agents deployed
One agent per asset across up to 50 workstations, servers, or cloud instances. Deployment is scripted against your existing management tooling; nothing needs a maintenance window.
- Days 1–3
Baseline and blind spots
The platform maps identities, endpoints, secrets, and external surface, then reports what it found that your current stack does not show — usually stale privilege and unmanaged assets.
- Days 4–10
Agents in recommend-only mode
Every decision the agent core would have executed is logged with its reasoning, so you can audit the verdicts before anything acts on your estate.
- Days 11–14
Autonomy, on your boundary
You widen the envelope where the recommendations earned it. We run a controlled adversary simulation against the scoped assets and you watch containment execute.
- Close
Findings, not a sales deck
A written report: what was found, what was contained, the measured time from detection to containment on your own estate, and what a production deployment would cost.
Start the deployment
A systems engineer picks this up within one business day and schedules the day-zero deployment. No sales representative attends unless you ask for one.
Before you start
Is this a limited trial build?
No. It is the production platform, capped at for 14 days, up to 50 assets, with the agents in recommend-only mode until you widen the envelope yourself.
What do you need from us to start?
A named technical owner, deployment access for the assets in scope, and read access to one identity provider. Everything else the agents discover themselves.
What happens to our data afterwards?
Telemetry collected during the evaluation is deleted within 30 days of the close report unless you convert, in which case it carries into your tenant. Nothing is used to train models shared across customers.
Can we run it against production?
Yes, and we recommend it — recommend-only mode exists so the first week carries no operational risk. Air-gapped and VPC deployments are available if telemetry may not leave your boundary.