outbound-sdr

name	outbound-sdr
description	Use when the user wants to build an outbound SDR campaign, create a target account list, research a market from ICP, find buyers, build account/persona POVs, or draft a 3-5 step outbound sequence. Runs a staged campaign workflow: ICP/value prop/offer intake, 3x long-list, preliminary scan, top-50% deep account research, final prioritized list, top-20 company POVs, buyer persona research, persona POVs, offer mapping, then message sequence.
user-invocable	true

Outbound SDR Engine

Purpose

Build a simplified Pipeline Rebel-style outbound workflow inside SalesSidekick Lite. The system should help a founder turn ICP, value proposition, and a clear campaign offer into a researched outbound motion.

This is not a "write me a cold email" shortcut. Messaging comes last.

Non-Negotiable Rule

Do not draft campaign messaging until the workflow has:

1. captured ICP, value proposition, and the specific campaign offer
2. created a long list at least 3x the final target size
3. completed a preliminary scan across the full long list
4. selected the top 50% for deeper account research
5. produced the final prioritized account list
6. created deep company POVs for the top 20 accounts
7. mapped buyer personas and researched individual buyers
8. created persona-level POVs
9. mapped the offer to each persona

If the user asks for emails too early, say what research gate is missing and offer to run the next gate.

Required Campaign Inputs

Collect these before Stage 1:

• ICP: industry, company size, geography, maturity, buying triggers, disqualifiers
• Value proposition: what the company sells, the business pain, and the outcome promised
• Clear offer: the campaign CTA or value asset that can earn a response
• Proof: customer examples, third-party validation, press, case studies, metrics, founder credibility, or useful content
• Final target size: if missing, default to 40 final target accounts
• Buyer personas: if missing, infer likely buying roles during account research

The clear offer can be a meeting ask, a useful playbook, a white paper, a benchmark, a teardown, a diagnostic, a customer proof point, a press announcement, or another asset that gives the prospect a reason to respond.

Connector Benefits

• Apollo: best source for account discovery, firmographic filtering, persona mapping, titles, emails, and contact enrichment.
• CRM: prevents duplicate outreach, respects current pipeline state, and surfaces existing account history.
• Search/Tavily: improves breadth and freshness of account discovery, news, hiring, trigger, press, and competitive research.
• Scraping/retrieval tools: useful for company sites, press pages, job posts, LinkedIn-visible context, and buyer profile pages when allowed.
• Email, calendar, Drive, call recordings: add first-party context, prior conversations, follow-up history, transcripts, and company assets.

If a connector is unavailable, use exports: Apollo CSV, CRM view export, call transcript folder, Drive docs, or pasted source material. Missing connectors reduce convenience, not the workflow.

Required Paths

Read:

• knowledge-bases/company-context.md
• data/setup-status.md
• data/connectors.md
• relevant files in data/research/
• relevant files in data/closeouts/
• relevant Slack/Teams digests if they contain account or buyer context

Write:

• data/outbound/{campaign-slug}/
• stage files under that campaign folder
• append one line to data/ops.log

Tool Requirements

• Apollo is preferred for account discovery, filters, contacts, titles, and enrichment. If unavailable, ask for Apollo CSV exports or user-provided lists.
• Search/Tavily or equivalent search is preferred for fresh account signals, press, hiring, category lists, and cited research. If unavailable, ask for pasted sources.
• CRM is preferred for exclusion checks and existing account context. If unavailable, ask for CRM CSV/Markdown exports.
• Drive/docs and call transcripts are preferred for proof assets and first-party context. If unavailable, ask for files or pasted excerpts.
• Do not scrape, enrich, or contact people through tools that are unavailable or unauthorized.
• Do not send email or push records to external systems without explicit approval.

Default Sizing

If the user does not specify a list size:

• Final target account list: 40 accounts
• Long list: at least 120 accounts
• Deep account research pool: top 60 accounts after preliminary scan
• Deep company POV pool: top 20 accounts

Never create a final list from fewer than 3x candidates unless the total addressable universe is genuinely smaller. If that happens, flag the limitation.

Stage 0: Campaign Intake

Ask only for missing essentials:

• What do we sell?
• Who is the ICP?
• Who is explicitly not a fit?
• What is the value proposition?
• What is the campaign offer or CTA?
• What proof or asset supports the offer?
• How many final target accounts do we want?
• Are Apollo, CRM, search/Tavily, and call transcript sources available?

Write the campaign plan to:

• data/outbound/{campaign-slug}/campaign-plan.md

Stage 1: Build The Long List

Create a long list at least 3x the final target size.

Use available sources in this order:

1. Apollo/account database
2. CRM exports and existing account lists
3. Search/Tavily category discovery
4. Relevant directories, partner pages, analyst lists, job posts, press releases
5. User-provided CSVs or pasted lists

Each account row should include:

• company name
• website
• segment/category
• location if relevant
• employee/revenue estimate if available
• why it might fit
• source
• disqualifier risk
• confidence

Save to:

• data/outbound/{campaign-slug}/01-long-list.md

Stage 2: Preliminary Scan And Stack Rank

Run a fast scan across the entire long list. Score every account.

Recommended scoring:

• ICP fit: 0-5
• trigger/timing signal: 0-5
• pain signal: 0-5
• reachable buyer personas: 0-5
• proof/offer relevance: 0-5
• evidence confidence: 0-5
• disqualifier risk: 0-5 negative

Output:

• total score
• priority tier
• reason for rank
• keep/deprioritize decision
• open research question

Take only the top 50% into deeper account research.

Save to:

• data/outbound/{campaign-slug}/02-preliminary-scan.md

Stage 3: Deep Account Research

Run deeper research on the top 50% from Stage 2.

For each account, research:

• current business context
• recent news, press, hiring, partnerships, funding, product launches, or strategic shifts
• likely pains and business pressures
• language the company uses to describe itself
• relevant buyer departments and likely owner
• why our value proposition may matter now
• sources and evidence quality

Create the final prioritized account list from this research.

Save to:

• data/outbound/{campaign-slug}/03-deep-account-research.md
• data/outbound/{campaign-slug}/04-final-prioritized-list.md

Stage 4: Top-20 Company POVs

For the top 20 accounts, create a serious account POV before researching individuals.

Each company POV should include:

• what is happening at the company
• why now might matter
• likely pain or opportunity
• what they probably care about commercially
• why the offer could be relevant
• a specific point of view
• sources used
• confidence and open questions

Save each POV to:

• data/outbound/{campaign-slug}/accounts/{company-slug}-pov.md

Stage 5: Buyer Persona Mapping

For each top account, identify buyer personas by buying role, not just title:

• authority
• budget owner
• pain owner
• blocker
• unblocker
• champion
• user

Use Apollo when available for names, titles, emails, LinkedIn URLs, seniority, departments, and contact confidence. If Apollo is missing, infer likely titles and ask for an export or pasted contacts.

Save buyer map to:

• data/outbound/{campaign-slug}/accounts/{company-slug}-buyer-map.md

Stage 6: Individual Buyer Research

Research each priority buyer deeply before writing to them.

Use available and permitted sources:

• Apollo contact data
• LinkedIn-visible profile context
• company bio pages
• podcasts, webinars, interviews, posts, quotes, and press
• CRM notes
• prior email or call history

For each person, capture:

• role and scope
• likely priorities
• what they are measured on
• likely pain
• relevance to the campaign offer
• personal/contextual hook
• what not to say
• source links or source notes
• confidence

Save to:

• data/outbound/{campaign-slug}/personas/{company-slug}/{person-slug}.md

Stage 7: Persona POV And Offer Mapping

Create a point of view for each buyer persona.

The persona POV must connect:

• company context
• person context
• likely business pressure
• the seller's value proposition
• the campaign offer
• a clear CTA

If the offer does not fit the persona, say so. Do not force a weak offer.

Stage 8: Sequence Drafting

Only after Stage 7, draft a 3-5 step sequence for each target person.

Each sequence should include:

• subject line
• email body or channel step
• specific account/person context
• clear offer or CTA
• proof or asset when relevant
• follow-up logic
• evidence notes

Rules:

• keep each email concise
• use the prospect's language when available
• do not cite unsupported facts
• do not over-personalize with creepy details
• do not send anything automatically

Save to:

• data/outbound/{campaign-slug}/sequences/{company-slug}-{person-slug}.md

Output Contract

At each stage, return:

• what was completed
• what was learned
• what was saved
• what is blocked or missing
• the next recommended stage

End every write with:

FILES UPDATED
- [path]

Append one line to data/ops.log for every saved stage.