AI for Sales, Outreach & Personalization — Prompt Library

Why this prompt library matters

One person can reach many more prospects safely and respectfully when AI handles repetitive research and drafting — but only if you keep humans in the loop, avoid over‑personalization, and follow simple privacy guardrails. This library gives practical prompts, scoring templates, outreach sequences, and review checklists you can copy, adapt, and test.

How to use these prompts

Copy a prompt, replace the placeholder fields (like {{prospect_name}}, {{company}}, {{pain_point}}), and run it against your preferred AI assistant. Always include a human review step before sending. Start with conservative personalization and iterate based on response rates and safety checks.

Persona & input fields to capture

Standardize the inputs your prompts expect so results stay consistent:

• prospect_name — full name
• role — job title
• company — company name
• industry — industry or niche
• relevant_public_info — neutral, public facts (funding, recent blog, product launch)
• observed_pain — short statement of a likely problem the prospect faces
• mutual_connection — if applicable
• product_offer — one‑line description of what you’re offering

Personalization prompts (examples)

Use conservative personalization: rely on public, professional facts; avoid scraping private data. The prompts below instruct the AI to be concise, professional, and to surface content for human editing.

Cold email — short opener

Prompt:

Draft a concise professional email opener (3–4 short sentences) to {{prospect_name}} at {{company}}. Use only public, business‑relevant facts from {{relevant_public_info}}. State one clear value idea based on {{observed_pain}}. End with a low‑friction question. Keep tone warm and human; avoid exaggeration and jargon.

LinkedIn connection + message

Write a 70–120 character connection note for {{prospect_name}} that references a public item {{relevant_public_info}} and briefly hints at a conversation about {{observed_pain}}. Do not include any private or personal information. Include a polite one‑sentence follow up suggestion after a connection is accepted.

Proposal intro paragraph

Create a one‑paragraph proposal intro that restates the prospect's likely challenge ({{observed_pain}}), explains how {{product_offer}} helps, and lists 2 measurable outcomes the prospect could expect. Keep it factual and conservative.

Lead scoring: simple rule templates

Start with transparent, easy rules you can automate. Assign points and prioritize totals.

• Industry fit: industry in target list = 20 pts
• Role fit: title matches buyer persona = 15 pts
• Engagement: visited pricing or demo page in last 90 days = 25 pts
• Company size: target range = 10 pts
• Signal strength: funding event or product launch in last 6 months = 20 pts
• Negative signal: competitor customer, not a buyer = -30 pts

Score interpretation (example): 60+ = high priority, 35–59 = nurture, <35 = low priority. Customize weights to reflect your funnel and test.

Email sequences with human‑in‑the‑loop checks

Sequence goals: 1) establish relevance, 2) offer clear value, 3) invite low‑effort next step. Always review AI drafts before sending.

1. Day 0 — Intro: Use the cold email opener prompt. Human check: confirm facts, remove any risky personalization, ensure subject line is truthful.
2. Day 3 — Reminder + value: Short note referencing prior message and adding one practical tip or link. Human check: ensure tip is genuinely useful and not misleading.
3. Day 8 — Case/example: Share a micro case study (1–2 sentences) showing measurable outcome for a similar client. Human check: verify accuracy and permissions for any customer names or stats.
4. Day 15 — Breakup: Polite final note offering to reconnect later. Human check: verify tone and opt‑out language.

Human review checklist (use before sending any AI‑generated outreach)

• All facts in the message are public and verifiable.
• No private/personal/sensitive information is referenced.
• Tone is respectful, not presumptive or manipulative.
• Promises or claims are supportable; no fabricated metrics.
• Clear next step is offered and easy to accept (e.g., 15‑minute call).
• Unsubscribe or opt‑out option is clear when sending sequences.

Privacy & ethical guardrails

Follow these simple rules to reduce risk and preserve reputation:

• Do not use nonpublic personal data (private email scraped from non‑professional sources, personal phone numbers, family details).
• Prefer company and role information that’s publicly available (LinkedIn public profile, press, company site).
• Limit behavioral tracking: only use signals you can justify (site visits, explicit downloads, form submissions).
• If using third‑party enrichment data, ensure the vendor’s data is compliant with relevant laws and your own privacy policy.
• Keep a record of what inputs were used to generate outreach (for transparency and later review).

Prompt engineering tips

• Be explicit about length, tone, and which inputs to use.
• Ask the model to provide a short edit‑ready version plus 2 alternative subject lines.
• Use temperature=0.2–0.6 for outreach drafts: lower for factual accuracy, slightly higher for warmth.
• Include a final instruction: “Return the email only; do not include any private data not in the inputs.”

Testing & metrics

Track simple metrics to learn fast: open rate, response rate, positive response rate (qualified replies), and conversion to meeting/demo. Run A/B tests on subject lines, first sentences, and value statements. Iterate weekly.

Next steps & recommended platform enhancements

Consider adding an interactive template that lets you paste prospect fields and returns ready‑to‑review drafts (human review required). For teams, bundle this library into a toolkit with lead scoring rules, saved templates, and an audit trail of inputs used for each outreach.

Copy and adapt these prompts to your voice and market. Start conservatively, record what you test, and continuously refine both prompts and human review steps.