Guardrails
Rules the system enforces, every time.
AI can write fast. These rules make sure it writes only what's true, only to people who should hear from you, and only from accounts that stay healthy. They're built into the system, not left to a prompt.
Every message
Nothing goes out that a person didn't mean.
A person approves every message
Drafts wait in AI Admin until someone approves, edits or skips them. You can let a single sequence send its own messages; answers to replies always wait for a person.
Only facts you gave it
The AI writes from your brand voice, knowledge base and company news, nothing else. No invented numbers, clients, results or quotes; if it doesn't know, it says it will confirm.
Drafts with a problem are flagged
A placeholder left in, a note over LinkedIn's length limit, or words that make a message sound AI-written are marked before you approve it.
Your tone, their style
Each message matches how the person likes to be written to. The tone changes; the facts never do.
The people you contact
Respect is built in.
Opt-outs are absolute
A clear "stop" or "unsubscribe" is caught before any AI reads the reply, and every other way of saying no is caught by the reply triage. Every sequence stops for that person, on every channel.
Consent where the law asks for it
WhatsApp and SMS go only to people who agreed, and the consent is recorded on each lead.
AI calls say they're AI
The AI caller rings only in its calling hours and tells people it's an AI.
Their working hours
Messages go out in each lead's own time zone and working hours.
An address and an opt-out
Your postal address and how to opt out go under every outreach email, as anti-spam rules require.
Your accounts
Sending that protects your reputation.
Daily limits
At most 100 emails a day per connected mailbox, 15 per bought inbox and 50 LinkedIn actions per account, with a gradual warm-up for new accounts.
LinkedIn slows down when it should
When an account needs care, new LinkedIn starts slow down; when it's at risk, they pause.
Clean lists
Email addresses are checked; bounced and invalid ones are never emailed again.
Healthy domains
SPF, DKIM and DMARC are checked for every sending domain, with what to fix when something is wrong.
Your money
No surprise bills.
The most it can cost, first
Every paid search or lookup shows the most it can cost before it runs.
Budgets that hold
The AI SDR agent stays within the weekly credit budget you set, and does fewer leads when it's tight.
Caps per client
Agencies give each client a monthly credit cap, so one client can never use up another's budget.
Who can do what
Access you control.
Roles
Owners, admins, members and viewers; view-only people can't change anything.
Client permissions enforced by the system
An agency's clients do only what the agency allowed. The server refuses the rest, not just the menu.
Keys that can't send
API keys and AI assistants can add and update leads, but never approve or send a message.
Signed webhooks
Every event we send carries a signature, so your systems know it came from us.
How we look after your data
Workspaces kept apart, encryption, exports and erasure on request: read trust and privacy and our privacy policy.
Put your outreach on autopilot, with you at the wheel.
Start free for 14 days. Connect an inbox or LinkedIn, describe who you sell to, and approve your first drafts today.
No card needed for the trial.