What we check before anything leaves the system — and what we do not yet check.
Enforced today
Enforced
Email (single, templated and campaign)
Every message is checked against the full compliance rule set before it reaches the mail provider. There is no setting that turns this off.
Enforced
Personal data in AI prompts
Social security numbers, card numbers, driver licence and health-insurance identifiers, emails and phone numbers are stripped before any text reaches an external model, and again on the way back.
Not yet covered
We would rather list these than let you assume they are handled.
Not yet
SMS
Not content-checked yet. Volume and consent controls apply.
Not yet
Generated documents
Not content-checked yet.
What blocks, and what only flags
On an outbound surface, anti-discrimination and Fair Housing rules refuse the send. You get the message back with the rule that stopped it, and you edit and retry. A false positive costs you one edit; a false negative is a Fair Housing violation on the record, so we hold the strict side of that trade.Inside the assistant, the same rules only flag. Words like “children” and “family status” come up constantly in legitimate conversation about who is buying a house, and blocking you mid-draft would be wrong. The check that matters runs at the point content would reach a consumer, not while you are thinking.
What we do not claim
We do not claim to guarantee Fair Housing compliance. No automated check can. These rules catch known patterns in the content we send on your behalf; they do not review your listings, your conversations off-platform, or your intent, and they are not legal advice.Compliance is not sold as an add-on. It is part of the product for every workspace, at every tier, with no upgrade path attached to it.