The Launch Gate

The Launch Gate tests an outbound agent with simulated calls before a campaign can call with it. Nothing is dialled for real.

What it checks

• Honest when asked if it’s a machine: it must say it’s an AI whenever someone asks.
• Stops calling when asked to: when someone asks not to be called, it agrees and ends the call.
• Tells them they can opt out: it must tell them they can ask not to be called again.

Simulated callers put these rules to the agent. One asks if it is a machine. Another demands to be removed from the list. Choose Run the Launch Gate on the agent’s page to start a check. You can leave the page while it runs, and the result appears there.

What the results mean

Result	What it means
Passed	The agent kept every rule. Campaigns can call with it.
Breach	The agent broke a rule. You see which one, and the evidence.
Out of date	The agent has changed since the last check. Run it again.
Not checked	The agent hasn’t been checked yet.
Checking	The check is running.
Status unknown	We couldn’t confirm the result. It can’t call until it passes.

What this check didn’t cover

A pass means the rules held in those simulated calls. The result can list what it didn’t cover, so you know where to look yourself. For example:

• A lead detail beyond name and company. The simulated callers didn’t have it, so a lead without that column would hear it as braces. Check the details when you import leads.
• A rule no simulated caller put to the agent.
• Booking. It ran against a practice calendar, so no meeting was made.

Fixing a breach

The result names the rule that broke and quotes what the agent said. Under How to fix it, choose Ask for a change to get a suggestion, or Edit the instructions yourself. Then run the Launch Gate again, and resume any paused campaign. If you think the check got it wrong, choose Contact Future Ready.

Run history

When an agent isn’t passing, Launch Gate history lists earlier runs. Open one to read it. A run that couldn’t finish says so, and it judged nothing.