Docs/Agents
Voice & Drafted Replies
Every draft is written in the voice of the campaign it belongs to: who's speaking, tone, your rules and samples, and the lines it never crosses. The words your whole company avoids live in Never say.
LeadPinger drafts a lot on your behalf: comments, DMs, replies, LinkedIn connection notes, the DM steps in sequences. All of it is written in the voice of the campaign the lead belongs to, and what goes out without you is up to that campaign's mode. Each campaign owns its whole voice, so a company with several offers runs one campaign per offer, each sounding the way that offer should. There is no workspace-wide voice page.
Where the voice lives
Open a campaign and go to its How your agent behaves step. Everything that shapes how it sounds is there:
| Field | What it decides |
|---|---|
| What you believe | The stance your agent holds in every reply, even the ones that never name the product. |
| Who's speaking | The person or role your agent writes as, in a sentence or two. Written as a role, it holds for every teammate who sends on this campaign. |
| Tone | Professional, Friendly, Casual, Direct or Enthusiastic: the attitude of every reply and message. |
| Language | The language every reply and message is written in. Empty means English. |
| Length | How long a public comment runs: One line, Short or Detailed. |
| Your rules | The specifics that make the text yours, like "No emoji, ever". Your rules win over the agent's house rules. |
| Your writing samples | 2 to 4 replies you wrote yourself. The strongest lever of all: the agent copies your rhythm, not your content. |
| This campaign never says | Words and phrases only this campaign avoids, on top of your company's Never say list. |
Product mentions follow the goal
How your product gets mentioned is not a separate setting: the campaign's goal decides it. Your agent names the product only when someone asks for recommendations and the community allows it. On the Get the product known goal it recommends the product from experience; every other goal names it once, in passing.
Hear it first
At the end of the step, Hear it first writes a real draft from your unsaved settings. Pick LinkedIn, Reddit or X and click Preview on a real post: it answers a post the campaign's agents actually found, or a sample post, marked as one, when they have found nothing there yet. Nothing is queued or sent. Change a field, click Try another post, and hear the difference before a lead does.
Never say
One list for the whole company lives in Settings, under Never say. Admins edit it, every campaign follows it on every platform, and it only ever rules things out: it never changes how a campaign sounds. It takes two kinds of entry:
- Phrase: exact words your company never says, like "guaranteed results", blocked word for word whatever the capitalization.
- Rule: a plain instruction every draft is given, like "Never quote a price".
Phrases are checked in every draft the AI writes. A draft that still uses one after a rewrite is held for your review and never goes out on its own, and a machine-approved message that breaks a phrase you added later goes back to approval before it sends. Anything you write and send yourself is sent as written.
Every campaign shows the company list, locked, just above its own This campaign never says, so nothing company-wide shapes a campaign invisibly.
What makes outreach actually get replies
- Reference their words.The drafts quote the lead's own problem back because that is what earns a reply. Keep that when you edit; strip the pitch, not the empathy.
- Samples beat adjectives. A tone is a hint; a few real replies you wrote are evidence. Paste them before you tune anything else.
- Fresh beats perfect. A good-enough reply while the post is hours old outperforms a perfect one two days later. This is why drafting is automatic and sending is one click.
Editing and regenerating
Every draft is editable in place, in the lead's Outreach tab and in Engage. Regenerate produces a fresh take in the same voice. Your edits are autosaved, and channel character limits are enforced as you type (LinkedIn connection notes cap at 300 characters), so nothing bounces at send time.
Questions this page didn't answer? Email support@leadpinger.ai, we usually reply within one business day.