Skip to main content

Training Center & message approvals

Control how the AI writes for each campaign with the Training Center, and review AI-generated messages in the approval queue before they're sent.

Two safety layers keep AI outreach on-brand: the Training Center defines how the AI writes (greetings, tone, rules), and message approvals let you review each AI-generated message before it's sent.

Where to find it

  • Per campaign: open the campaign and select the Training Center tab.

  • Personal defaults: Settings → Personal Settings → Messaging Style.

  • Company context for the AI: Settings → Account Settings → Training Center.

  • The approval queue: Outreach → Activities.

The campaign Training Center

The Training Center tab shows three cards. Each is Inherited from the campaign owner's personal Messaging Style by default — click Override on a card to set campaign-specific values, or Reset to default to fall back to the profile again:

  • Greetings & Signatures — the Default Salutation (e.g. "Hi") and Default Ending (sign-off) per language. The AI picks the one matching each message's language.

  • Tone of Voice — a short description of how messages should sound, per language (e.g. "professionally informal, personal, and free from generic recruiter language").

  • Message Rules — two parts: System rules (always enforced) — NoRecruit's built-in guardrails against recruiter clichés, buzzwords, fake links, and passive closings — shown read-only, and Your custom rules (e.g. "Always mention the remote work option").

Changes appear in a You have unsaved changes bar — click Save Changes to apply them to the campaign.

💡 Tip: Set your defaults once under Settings → Personal Settings → Messaging Style — every campaign you own inherits them automatically. Use campaign overrides only where a role genuinely needs a different voice. Salutation and tone of voice are required before a campaign can launch.

At team level, Settings → Account Settings → Training Center stores your company context — Company Name, Website, and About Us — which the AI uses as background when writing messages.

Turning on message approvals

On any AI messaging block in the Campaign Builder, the Requires Approval toggle controls whether "AI-generated messages require manual approval before sending". With it on, each generated message is created with status Pending Approval and waits for you instead of going out automatically.

You can set this per block — for example, require approval on the first touch but let short follow-ups send automatically.

Reviewing and approving messages

  1. Go to Outreach → Activities.

  2. Filter by status Pending Approval (there's also a stats tile for it).

  3. Open an activity to review the AI-generated message. From the Edit Outreach Message dialog you can:

    • Edit the text (and subject, where applicable) directly, then Save Message.

    • Regenerate with AI — optionally give feedback first ("Make it more formal", "Add more enthusiasm") and the AI rewrites the message.

    • View original AI-generated content if you've edited it.

  4. Click Approve — the message is scheduled and sent within your working hours.

Approving in bulk

Select multiple activities in the list and use Approve Activities to approve them at once. You can also bulk Regenerate Messages (with shared feedback) or Pause Activities. Note that activities from a stopped campaign can't be approved or scheduled.

💡 Tip: Approvals are a great way to train the AI in your first weeks. Once the messages consistently sound right, switch Requires Approval off for follow-up blocks and keep it on for first touches only.

FAQ

Whose Messaging Style does a campaign inherit? The campaign owner's — messages are sent from the owner's accounts, so the AI writes in their voice. Other team members editing the campaign see the owner's inherited values.

Do approved messages send immediately? They're scheduled into the sending account's working hours with natural spacing, like all other outreach.

What happens if I never approve a message? It simply stays in Pending Approval — nothing is sent for that step, and the profile waits at that point in the sequence.

Did this answer your question?