Sales Prospecting Sequence

Prospecting Sequence Generation

Standalone sales workflow for one-to-one B2B outbound email. It is separate from marketing content drafts and should not use blog/social prompt layers.

When to Use

• Outbound email sequence for a service + ICP + persona
• Idea with content_type: prospecting_sequence OR direct user request
• Founder-led or sales-led outreach where the goal is a reply, diagnosis, or

low-friction next step

Do not use for newsletters, nurture campaigns, product onboarding, blog/social drafts, or mass promotional sends.

Prerequisites

• workspace/firm/profile.md
• workspace/firm/services/{service}.md
• workspace/marketing/icp/{icp}.md
• workspace/marketing/icp/personas/{persona}.md
• Current agent/LLM session
• Optional content idea with content_type: prospecting_sequence
• Optional account/prospect facts supplied by the user

If firm, service, ICP, or persona context is missing, stop and create it first. Do not invent an inline ICP/persona inside this skill.

No Exa research by default. Use only workspace facts and user-provided account facts. If account-specific personalization is requested but facts are missing, write placeholders and list the missing facts instead of fabricating triggers.

Workflow

1. Load context and choose mode

Read the firm, service, ICP, and persona files. If an idea file exists, read all frontmatter, especially language, buyer_question, unique_angle, proof_source, and next_action.

Choose one mode:

• segment_sequence - reusable sequence for a service + ICP + persona.
• account_sequence - sequence for one named account/prospect with supplied

facts.

• variant_pack - base sequence plus 2-3 light variants for different personas

or trigger situations.

2. Build the strategy

Define before writing emails:

• audience: target company type + buyer persona;
• offer: the useful thing the prospect can get before a sales call;
• proof: only facts from workspace/user input, or mark proof_needed;
• likely objection: time, risk, internal capacity, incumbent vendor, budget;
• CTA type: diagnostic question, useful resource, teardown/audit offer, or soft

permission ask;

• personalization mode: segment, account, or placeholder.

Do not open with "we help companies like yours" unless the segment and problem are concrete. Prefer a specific observation, tension, trigger, or buyer problem.

3. Select one framework

The model picks ONE framework from ICP/persona signals:

| Framework | Select when | |-----------|-------------| | ois (Observation-Insight-Solution) | ICP has concrete industry, companysize, techstack — specific data to observe | | pas (Problem-Agitation-Solution) | ICP/persona have strong, clearly defined painpoints + fearsrisks | | trigger (Trigger-Based) | Persona has triggerevents (funding, hiring, expansion) | | value_first | Persona seniority director+, changereadiness cautious, or aiattitude neutral/cautious/skeptic | | curiosity (Curiosity-Gap) | Persona senioritylevel is c_level or vp |

4. Choose email count and cadence

Start from ICP sales_cycle_length, then adjust for relationship warmth:

| salescyclelength | Emails | |--------------------|--------| | 1_week, 1_month | 3 (short cycle) | | 1_3_months | 5 (standard B2B) | | 3_6_months, 6_12_months, over_1_year | 7 (enterprise) |

Typical email-only spacing: Day 0 -> 3 -> 7 -> 12 -> 18 -> 25 -> 35.

Rules:

• 3-5 emails is the default for cold outbound.
• Use 7 emails only for high-value enterprise sequences with a real reason to

keep adding new angles.

• Each follow-up must add a new angle: different pain, proof, useful asset,

objection, trigger, or breakup. Never write "just following up".

• Stop the sequence on reply, opt-out, bounce, or clear disinterest.

5. Write the emails

Each email:

• 50-125 words by default; 150 words max when context is complex.
• Plain text only. No markdown bullets inside email bodies.
• One idea, one CTA, one expected reply.
• CTA should be a low-friction offer or question before asking for a meeting.
• Subject line must be honest and specific, never clickbait.
• Mention proof only if it exists in workspace/user input.
• Include personalization placeholders only when needed:

{prospect_company}, {prospect_role}, {observed_trigger}, {relevant_system}, {sender_name}, {sender_company}, {postal_address}, {opt_out_line}.

All output must be written in the idea's or user's requested language, and must use the firm's brand voice + respect words_to_avoid.

6. Run the quality gate

Before writing the final file, check:

• Is the prospect problem specific to the ICP/persona?
• Does every email add a new reason to reply?
• Is every claim/proof sourced from workspace or marked proof_needed?
• Is the CTA smaller than "book 30 minutes" unless the prospect is warm?
• Are fake triggers, fake case studies, fake metrics, and fake client names absent?
• Are opt-out/sender identity placeholders present for commercial outreach?
• Is the sequence suitable for one-to-one sales outreach, not mass marketing?

This skill does not provide legal advice. The output should include a compliance checklist and remind the sender to verify local law, consent/legitimate-interest rules, unsubscribe handling, and sender authentication before sending.

Output Rules

Use the templates in references/output-templates.md.

Default output is a single Markdown file:

workspace/sales/prospecting/{service}--{icp}--{persona}--{campaign-slug}.md

Use an output folder only when the user asks for variants, CRM import, or account customization pack:

workspace/sales/prospecting/{service}--{icp}--{persona}--{campaign-slug}/
├── sequence.md              # canonical ready-to-send sequence
├── variants.md              # optional persona/trigger variants
└── crm-import.csv           # optional flat export, generated only if requested

Keep the filename/folder self-describing because prospecting can produce many similar artifacts over time.

Internal Structured Draft

{
  "selected_framework": "ois|pas|trigger|value_first|curiosity",
  "framework_reasoning": "string",
  "mode": "segment_sequence|account_sequence|variant_pack",
  "personalization_mode": "segment|account|placeholder",
  "email_count": 3,
  "cadence_days": [0, 3, 7],
  "offer": "string",
  "proof_status": "workspace_proof|user_supplied|proof_needed",
  "sequence": [
    {
      "email_number": 1,
      "subject_line": "string",
      "body": "string (50-125 words)",
      "purpose": "initial_contact|follow_up|value_add|social_proof|urgency|breakup",
      "cta_type": "diagnostic_question|resource_offer|audit_offer|permission_ask|meeting_ask|breakup",
      "days_after_previous": 0
    }
  ],
  "quality_gate": {
    "claims_sourced": true,
    "fake_triggers_absent": true,
    "one_cta_per_email": true,
    "opt_out_placeholder_present": true
  }
}

This is an internal planning shape, not a separate JSON deliverable unless the user explicitly asks for JSON.

References

• references/output-templates.md - required Markdown structure, optional

variants and CSV export templates

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: b2bforce
• Source: b2bforce/b2bforce
• License: MIT
• Homepage: https://www.b2bforce.ai/

Install and usage instructions live in the source repository linked above.