Trust & data

Your data, inside
the boundary you agree.

Your data is read only within the boundary we agree before deployment, and agents act on nothing without a person confirming.

Ask us the questions your IT team has

The boundary

The boundary is agreed
before we build.

What is read, where it runs, what never leaves — agreed up front, not discovered later.

  1. Define: What the system may read, where processing happens and what never leaves.
  2. Permission: Access mirrors your role-based permissions; nothing reads wider than the boundary.
  3. Log: What was read, produced and confirmed is traceable.

How we deploy AI

No confirmation, no action

Draft, propose, flag.
You confirm.

Agents never submit, send or publish on their own. Every output waits for a person.

Meet the agents

What you'll see

Interfaces shown
with de-identification.

Where we show a live interface, names, amounts and brands are replaced or rounded — and the caption tells you what's real and what's not.

See the proof

What we don't claim

What we have
not yet published.

On data residency, encryption, training-data use and certifications, we have not published specifics. Those will be added once the facts are confirmed — we'd rather leave them blank than state what we can't verify.

  • Data residency: Not yet published. It will be stated once the facts are confirmed.
  • Encryption: Not yet published. It will be stated once the facts are confirmed.
  • Training-data use: Not yet published. It will be stated once the facts are confirmed.
  • Certifications: None claimed. We state only what we can verify.

How we deploy AI

Bring it

Bring us your IT team's questions.

Data questions are best answered on your terms. Bring the list your IT or procurement team has.

Start a conversation · Data & methodology · Privacy policy