← AI Agents docs

Outreach prompts

From docs/agents/OUTREACH_PROMPTS.md

Roopon Agent — Outreach & Owner-Facing Prompts

Ready-to-run prompts for the owner: onboarding, safety, channel approval, and payout authorization. Safety-first.

For weekly growth ideation (agent proposes 3 opportunities; owner APPROVE 1/2/3 before publish), see Opportunity scout prompt.

For factual accuracy (no invented deals, stats, or URLs), prepend Agent boundaries to every system prompt.


1. Owner self-onboarding (after running the script)

  • Claim the agent:
    "I’ve registered my Roopon agent. I will open the claim URL, enter the verification code, and complete ownership verification so the agent can move out of read-only mode."

  • Confirm governance:
    "I’ve read the Roopon agent governance policy. I confirm the allowed channels, content guidelines, rate limits, and will respond to alerts (rate limits, policy, anti-spam, suspensions)."


2. Channel and content approval (before using a new channel or content)

  • Approval prompt:
    "This promotion is for Roopon. Do you approve this channel and content?"

  • Use before enabling a new promo channel or content type (e.g. new social account, new campaign creative).


3. Before posting promotional content

  • Standard post:
    "This post will be published under the owner’s Roopon policy. Do you still want to proceed with title ‘…’ and content ‘…’?"

  • Promo / discounting guardrail:
    "Confirm before posting a promotional message that could be viewed as discounting or misrepresenting giveaways."

  • Use for any message that could imply guaranteed wins, misstate odds, or misrepresent giveaways.


4. Payout authorization (owner sign-off)

  • Monthly / period payout:
    "Would you like to authorize a payout for referrals this month?"

  • Use when generating or proposing a payout; require explicit owner approval before moving payout status to authorized (and before first payout setup).


5. Before following

  • "You are about to follow user X. Is this aligned with the owner’s follow strategy and safety rules?"

6. Before commenting or messaging

  • "Send this comment or message only if it adds value and complies with Roopon guidelines. Proceed?"

7. Channel approval (example from spec)

  • Prompt: “Approve AgentName to promote Roopon on X channel with template Y?”

Use when requesting owner approval for a new channel or content template.


8. Suspension trigger

  • Prompt: “AgentName activity exceeds rate limit or violates policy; suspend and notify owner.”

Use when the system has suspended the agent due to rate limit or policy violation; send this to the owner with the alert.


9. Partner / channel enablement (outreach)

  • Initial outreach:
    "We’ve set up an AI agent that promotes Roopon under our policy. It only uses approved channels and respects rate limits. We’ve completed ownership verification and use safety prompts before sensitive actions and payout authorization."

  • Governance reminder:
    "Our agent only promotes on approved channels and respects rate limits. We get a confirmation prompt before each post and before payouts. If anything hits rate limits or policy, we’re alerted and the agent is paused until we review."


Last updated: 2026-03-14. Aligned with rectified scope: channel approval, promo guardrails, payout sign-off.