Trust
AI in the loop.
People own the result.
Agents write most of our code, copy, and campaigns. These are the rules for how that work gets reviewed, shipped, and owned, on partner projects and inside our own ventures.
People set the goal
Every task starts with a human-written brief and a named owner. Agents execute. They never decide what to build or who to target.
Review before release
Nothing generated ships unread. Code gets a senior review, copy gets an editor, and campaigns get a claims and budget check before spend.
Your data stays yours
Partner data, repos, and audiences never train a model. We use them to do the work you hired us for, then hand them back.
Six rules
What we hold ourselves to
Short enough to remember, specific enough to check. Every engagement is run against them.
People set the goal
Every task starts with a human-written brief and a named owner. Agents execute. They never decide what to build or who to target.
Review before release
Nothing generated ships unread. Code gets a senior review, copy gets an editor, and campaigns get a claims and budget check before spend.
Your data stays yours
Partner data, repos, and audiences never train a model. We use them to do the work you hired us for, then hand them back.
Say what is generated
Partners see which deliverables agents produced. Synthetic voices, faces, and footage are labeled wherever they could pass for real.
Every agent has an off switch
Deployed agents run with scoped credentials, spend limits, logs, and a person to escalate to. Any of them can be paused in one step.
Only work we would sign
We turn down deceptive ads, dark patterns, scraped audiences, and products built to mislead. If we ship it, our name is on it.
Commitments
How you can check
Rules are cheap. These are the artifacts that let a partner verify we follow them.
Public playbooks
Our engineering and marketing playbooks are published, including the review gates agent output must pass. Read them before you hire us.
An AI section in every scope
Each statement of work lists which tasks agents run, which tools they use, and who reviews the output before it reaches you.
Logs you can read
Agent runs on partner projects are logged and available in the client portal for the life of the engagement, and exported at handover.
A dated review of these rules
We revisit this page when tooling, law, or our own mistakes change what responsible use looks like. Every change lands in the changelog.
Reporting
Raise a concern
Seen something an Entatan agent, ad, or product did that it should not have? Tell us. Every report is read by a person and acknowledged within two business days.
Misleading ads, harmful outputs, or an agent acting outside its brief. Use the contact form and choose the trust topic.
Security vulnerabilitiesWeaknesses in a site, product, or venture. Follow the disclosure process on the security page.
Your personal dataAccess, correct, delete, or restrict how we use data about you through the privacy portal.
Help us hold the line
We hire engineers and marketers who will push back on a brief. If that sounds like you, look at the open roles.