The Complete Guide for
Sales & GTM 2027.

Navigation

Navigate the 12 Sales & GTM operating systems. Use the page ranges below to move directly to the dashboard, prompts, Skills, code, provider boundary, recovery and rollout.

Every system exposes the complete operating stack.

An independent guide by NipPro AI. A NipPro AI publication by Alain Nono, Co-Founder & CEO. This training is offered to you by NipPro AI. PDF pages 1–3. Open the original PDF (opens Google Drive in a new tab)

Operating Doctrine

Claude can run more of the GTM process - without owning the commercial judgment.

Separate what Claude can help produce from what only evidence, an authorized system or a human decision can establish.

The Claude stack used in this guide.

Capabilities are applied only where they improve an inspectable GTM job. Availability, models and syntax must be rechecked before implementation.

Primary Anthropic sources for Claude claims. Independent GTM operating models for the work.

Sources checked 18 October 2026. Product capabilities, URLs, models and availability can change; recheck before implementation.

Twelve operating systems. One evidence-preserving revenue architecture.

Every system shares typed state, but each owns distinct business objects, helpers, prompts, Skills, tools and recovery.

Security and decision doctrine.

No prompt, Skill, tool, connector or dashboard is authority to perform a consequential action.

PDF pages 4–8.

Global Practice

Fifteen global workspaces, PDF pages 9–23. Every global page carries the same operating contract:

Reader Workspace Start

Choose one of 12 systems, load a fixture and preserve the state.

workspace_id: demo-gtm-2027
view: workspace-start
linked_file: site/index.html
status: READY_FOR_REVIEW
executed_actions: []

Operating workspace: Source → Prompt → Skill → Code → QA

Inspector: Coverage 84% · Restrictions 2 active · Receipt None

Source-of-Truth Registry

Register product truth, buyer evidence, CRM state and policy sources before prompting.

Linked file: schemas/source_record.schema.json

GTM Data Contract

Define identity, evidence, dates, restrictions, actions, receipts and handoff fields.

Linked file: schemas/gtm_packet.schema.json

Master Prompt Compiler

Compile a source-bound operating prompt with output contract, tool plan and failure behavior.

# Master Prompt Compiler
ROLE: source-bound GTM operator
INPUT: workspace, sources, restrictions
DETERMINISTIC: run helper first
OUTPUT: typed packet + evidence IDs
BOUNDARY: proposal != execution
CHECK: facts, math, identity, receipt
RECOVERY: repair smallest dependency
FILE: prompts/00-master-gtm-operating-system.md

Validated: Inputs complete • schema attached • no hidden side effect

Global Skill Router

Route the dominant job to one primary Skill and keep adjacent Skills unloaded until needed.

Linked file: skills/00-gtm-system-router/SKILL.md

Tool Search & Deferred Loading Lab

Compare all-tools-loaded with router, Skills and Tool Search + defer_loading.

# Tool Search & Deferred Loading Lab
ROLE: source-bound GTM operator
INPUT: workspace, sources, restrictions
DETERMINISTIC: run helper first
OUTPUT: typed packet + evidence IDs
BOUNDARY: proposal != execution
CHECK: facts, math, identity, receipt
RECOVERY: repair smallest dependency
FILE: worksheets/tool_search_benchmark.json

Validated: Inputs complete • schema attached • no hidden side effect

Structured Output Lab

Use schema-constrained output, then validate facts and business rules separately.

# Structured Output Lab
ROLE: source-bound GTM operator
INPUT: workspace, sources, restrictions
DETERMINISTIC: run helper first
OUTPUT: typed packet + evidence IDs
BOUNDARY: proposal != execution
CHECK: facts, math, identity, receipt
RECOVERY: repair smallest dependency
FILE: schemas/handoff_packet.schema.json

Validated: Inputs complete • schema attached • no hidden side effect

Files & Source Packet Lab

Upload/reuse sources by reference while preserving workspace isolation and expiry.

Linked file: worksheets/files_source_packet.md

Prompt Caching Lab

Cache stable policy and source prefixes without hiding invalidation or stale evidence.

# Prompt Caching Lab
ROLE: source-bound GTM operator
INPUT: workspace, sources, restrictions
DETERMINISTIC: run helper first
OUTPUT: typed packet + evidence IDs
BOUNDARY: proposal != execution
CHECK: facts, math, identity, receipt
RECOVERY: repair smallest dependency
FILE: worksheets/prompt_caching_plan.md

Validated: Inputs complete • schema attached • no hidden side effect

Batch Eval Lab

Prepare asynchronous eval sets with custom IDs, results and reconciliation.

# Batch Eval Lab
ROLE: source-bound GTM operator
INPUT: workspace, sources, restrictions
DETERMINISTIC: run helper first
OUTPUT: typed packet + evidence IDs
BOUNDARY: proposal != execution
CHECK: facts, math, identity, receipt
RECOVERY: repair smallest dependency
FILE: evals/gtm_eval_tasks.json

Validated: Inputs complete • schema attached • no hidden side effect

Model, Effort & Token Planner

Select a current model, effort and token budget by complexity, latency and review need.

# Model, Effort & Token Planner
ROLE: source-bound GTM operator
INPUT: workspace, sources, restrictions
DETERMINISTIC: run helper first
OUTPUT: typed packet + evidence IDs
BOUNDARY: proposal != execution
CHECK: facts, math, identity, receipt
RECOVERY: repair smallest dependency
FILE: worksheets/model_effort_plan.csv

Validated: Inputs complete • schema attached • no hidden side effect

Connector & MCP Boundary

Separate messaging, data access, client tools, MCP, provider actions and receipts.

Linked file: worksheets/runtime_architecture.md

Security & Approval Lab

Protect secrets, isolate untrusted content and require approval for consequential actions.

Linked file: .env.example

Route & State Ledger

Trace identity, evidence, restrictions, proposals, receipts and recovery across systems.

workspace_id: demo-gtm-2027
view: route-ledger
linked_file: routes/gtm_automation_full_route.json
status: READY_FOR_REVIEW
executed_actions: []

Operating workspace: Source → Prompt → Skill → Code → QA

Inspector: Coverage 84% · Restrictions 2 active · Receipt None

Release Audit Dashboard

Verify title, counts, tests, sources, links, packaging and unresolved external gates.

workspace_id: demo-gtm-2027
view: release-audit
linked_file: MANIFEST.json
status: READY_FOR_REVIEW
executed_actions: []

Operating workspace: Source → Prompt → Skill → Code → QA

Inspector: Coverage 84% · Restrictions 2 active · Receipt None

PDF pages 9–23.

GTM Strategy & Planning

System 01 · GTM Command Center. Turn product truth, market evidence, constraints and revenue goals into a dated, reviewable GTM plan. PDF pp. 024-046.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/gtm-strategy/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: 48b00e34458047e0

Linked file: scripts/systems/01-gtm-strategy/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: large TAM with no proof

Linked file: skills/01-gtm-strategy-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: GTM Strategy & Planning research operator
JOB: Turn product truth, market evidence, constraints and revenue goals into a dated, reviewable GTM plan.
INPUTS: product truth, market segments, revenue goal, capacity
EVIDENCE: approved product truth, market evidence, customer evidence, capacity record
DETERMINISTIC: weighted segment priority, capacity utilization, budget guardrail
OUTPUT: market thesis, segment decision table, 90-day plan
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/01-gtm-strategy/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: GTM Strategy & Planning analyst operator
JOB: Turn product truth, market evidence, constraints and revenue goals into a dated, reviewable GTM plan.
INPUTS: product truth, market segments, revenue goal, capacity
EVIDENCE: approved product truth, market evidence, customer evidence, capacity record
DETERMINISTIC: weighted segment priority, capacity utilization, budget guardrail
OUTPUT: market thesis, segment decision table, 90-day plan
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/01-gtm-strategy/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: GTM Strategy & Planning artifact operator
JOB: Turn product truth, market evidence, constraints and revenue goals into a dated, reviewable GTM plan.
INPUTS: product truth, market segments, revenue goal, capacity
EVIDENCE: approved product truth, market evidence, customer evidence, capacity record
DETERMINISTIC: weighted segment priority, capacity utilization, budget guardrail
OUTPUT: market thesis, segment decision table, 90-day plan
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/01-gtm-strategy/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/01-gtm-strategy/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# gtm-strategy reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/01-gtm-strategy/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/gtm-strategy/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 024-046.

ICP Definition & Account Prioritisation

System 02 · ICP Prioritization Engine. Build an interpretable account queue from policy weights, disqualifiers, observed fit and evidence coverage. PDF pp. 047-069.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/icp-prioritization/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: 6a94dd6223a711e9

Linked file: scripts/systems/02-icp-prioritization/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: high score with unresolved exclusion

Linked file: skills/02-icp-prioritization-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: ICP Definition & Account Prioritisation research operator
JOB: Build an interpretable account queue from policy weights, disqualifiers, observed fit and evidence coverage.
INPUTS: ICP policy, accounts, evidence, weights
EVIDENCE: ICP policy version, authorized account source, firmographic record, hard-stop evidence
DETERMINISTIC: weighted fit, observed coverage, hard-stop precedence
OUTPUT: interpretable scorecard, coverage report, hard-stop result
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/02-icp-prioritization/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: ICP Definition & Account Prioritisation analyst operator
JOB: Build an interpretable account queue from policy weights, disqualifiers, observed fit and evidence coverage.
INPUTS: ICP policy, accounts, evidence, weights
EVIDENCE: ICP policy version, authorized account source, firmographic record, hard-stop evidence
DETERMINISTIC: weighted fit, observed coverage, hard-stop precedence
OUTPUT: interpretable scorecard, coverage report, hard-stop result
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/02-icp-prioritization/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: ICP Definition & Account Prioritisation artifact operator
JOB: Build an interpretable account queue from policy weights, disqualifiers, observed fit and evidence coverage.
INPUTS: ICP policy, accounts, evidence, weights
EVIDENCE: ICP policy version, authorized account source, firmographic record, hard-stop evidence
DETERMINISTIC: weighted fit, observed coverage, hard-stop precedence
OUTPUT: interpretable scorecard, coverage report, hard-stop result
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/02-icp-prioritization/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/02-icp-prioritization/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# icp-prioritization reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/02-icp-prioritization/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/icp-prioritization/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 047-069.

Prospect & Account Research

System 03 · Account Intelligence Workspace. Create sourced account and contact briefs without turning mentions, guesses or stale evidence into facts. PDF pp. 070-092.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/account-research/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: bb1d833d2b5b89fd

Linked file: scripts/systems/03-account-research/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: integration page mistaken for prospect stack

Linked file: skills/03-account-research-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: Prospect & Account Research research operator
JOB: Create sourced account and contact briefs without turning mentions, guesses or stale evidence into facts.
INPUTS: account IDs, authorized sources, research questions, freshness policy
EVIDENCE: authorized source list, captured passage, source timestamp, record locator
DETERMINISTIC: source freshness, dedupe, claim coverage
OUTPUT: account brief, contact brief, timeline
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/03-account-research/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: Prospect & Account Research analyst operator
JOB: Create sourced account and contact briefs without turning mentions, guesses or stale evidence into facts.
INPUTS: account IDs, authorized sources, research questions, freshness policy
EVIDENCE: authorized source list, captured passage, source timestamp, record locator
DETERMINISTIC: source freshness, dedupe, claim coverage
OUTPUT: account brief, contact brief, timeline
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/03-account-research/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: Prospect & Account Research artifact operator
JOB: Create sourced account and contact briefs without turning mentions, guesses or stale evidence into facts.
INPUTS: account IDs, authorized sources, research questions, freshness policy
EVIDENCE: authorized source list, captured passage, source timestamp, record locator
DETERMINISTIC: source freshness, dedupe, claim coverage
OUTPUT: account brief, contact brief, timeline
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/03-account-research/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/03-account-research/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# account-research reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/03-account-research/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/account-research/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 070-092.

Outbound Sequences & Personalisation

System 04 · Sequence Architecture Lab. Design evidence-based multi-touch sequences with clear stop rules, channel logic and review gates. PDF pp. 093-115.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/sequence-personalization/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: e4fb4e16e49297e3

Linked file: scripts/systems/04-sequence-personalization/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: same claim repeated across channels

Linked file: skills/04-sequence-personalization-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: Outbound Sequences & Personalisation research operator
JOB: Design evidence-based multi-touch sequences with clear stop rules, channel logic and review gates.
INPUTS: offer, audience, proof, channel constraints
EVIDENCE: approved offer claims, contact evidence, channel permission, preference state
DETERMINISTIC: cadence intervals, touch caps, cross-channel collision
OUTPUT: sequence map, touch drafts, personalization map
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/04-sequence-personalization/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: Outbound Sequences & Personalisation analyst operator
JOB: Design evidence-based multi-touch sequences with clear stop rules, channel logic and review gates.
INPUTS: offer, audience, proof, channel constraints
EVIDENCE: approved offer claims, contact evidence, channel permission, preference state
DETERMINISTIC: cadence intervals, touch caps, cross-channel collision
OUTPUT: sequence map, touch drafts, personalization map
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/04-sequence-personalization/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: Outbound Sequences & Personalisation artifact operator
JOB: Design evidence-based multi-touch sequences with clear stop rules, channel logic and review gates.
INPUTS: offer, audience, proof, channel constraints
EVIDENCE: approved offer claims, contact evidence, channel permission, preference state
DETERMINISTIC: cadence intervals, touch caps, cross-channel collision
OUTPUT: sequence map, touch drafts, personalization map
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/04-sequence-personalization/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/04-sequence-personalization/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# sequence-personalization reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/04-sequence-personalization/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/sequence-personalization/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 093-115.

Cold Email, LinkedIn & Multichannel Outreach

System 05 · Channel Orchestration Console. Coordinate reviewed outreach across email, LinkedIn and calls without confusing drafts with executed platform actions. PDF pp. 116-138.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/multichannel-outreach/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: 6c0a01e6e1b13241

Linked file: scripts/systems/05-multichannel-outreach/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: draft logged as sent

Linked file: skills/05-multichannel-outreach-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: Cold Email, LinkedIn & Multichannel Outreach research operator
JOB: Coordinate reviewed outreach across email, LinkedIn and calls without confusing drafts with executed platform actions.
INPUTS: approved contacts, channel permissions, message drafts, send windows
EVIDENCE: contact record, permission basis, approved draft, restriction record
DETERMINISTIC: send eligibility, cross-channel spacing, preference scope
OUTPUT: channel plan, draft packet, collision alert
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/05-multichannel-outreach/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: Cold Email, LinkedIn & Multichannel Outreach analyst operator
JOB: Coordinate reviewed outreach across email, LinkedIn and calls without confusing drafts with executed platform actions.
INPUTS: approved contacts, channel permissions, message drafts, send windows
EVIDENCE: contact record, permission basis, approved draft, restriction record
DETERMINISTIC: send eligibility, cross-channel spacing, preference scope
OUTPUT: channel plan, draft packet, collision alert
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/05-multichannel-outreach/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: Cold Email, LinkedIn & Multichannel Outreach artifact operator
JOB: Coordinate reviewed outreach across email, LinkedIn and calls without confusing drafts with executed platform actions.
INPUTS: approved contacts, channel permissions, message drafts, send windows
EVIDENCE: contact record, permission basis, approved draft, restriction record
DETERMINISTIC: send eligibility, cross-channel spacing, preference scope
OUTPUT: channel plan, draft packet, collision alert
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/05-multichannel-outreach/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/05-multichannel-outreach/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# multichannel-outreach reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/05-multichannel-outreach/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/multichannel-outreach/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 116-138.

Call Prep, Discovery & Objection Handling

System 06 · Conversation Intelligence Desk. Turn sourced account context into call preparation, discovery paths and bounded objection responses. PDF pp. 139-161.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/calls-discovery/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: dd6c829f54d66dc7

Linked file: scripts/systems/06-calls-discovery/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: invented urgency

Linked file: skills/06-calls-discovery-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: Call Prep, Discovery & Objection Handling research operator
JOB: Turn sourced account context into call preparation, discovery paths and bounded objection responses.
INPUTS: account brief, contact role, known context, product proof
EVIDENCE: account facts, contact role, product proof, call transcript or notes
DETERMINISTIC: agenda timing, question coverage, objection taxonomy
OUTPUT: call brief, discovery tree, objection response
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/06-calls-discovery/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: Call Prep, Discovery & Objection Handling analyst operator
JOB: Turn sourced account context into call preparation, discovery paths and bounded objection responses.
INPUTS: account brief, contact role, known context, product proof
EVIDENCE: account facts, contact role, product proof, call transcript or notes
DETERMINISTIC: agenda timing, question coverage, objection taxonomy
OUTPUT: call brief, discovery tree, objection response
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/06-calls-discovery/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: Call Prep, Discovery & Objection Handling artifact operator
JOB: Turn sourced account context into call preparation, discovery paths and bounded objection responses.
INPUTS: account brief, contact role, known context, product proof
EVIDENCE: account facts, contact role, product proof, call transcript or notes
DETERMINISTIC: agenda timing, question coverage, objection taxonomy
OUTPUT: call brief, discovery tree, objection response
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/06-calls-discovery/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/06-calls-discovery/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# calls-discovery reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/06-calls-discovery/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/calls-discovery/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 139-161.

Sales Collateral, Decks & Battlecards

System 07 · Sales Asset Studio. Build decks, one-pagers and battlecards from approved claims, current proof and versioned source material. PDF pp. 162-184.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/sales-collateral/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: c845ecb30c54834f

Linked file: scripts/systems/07-sales-collateral/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: old competitor claim

Linked file: skills/07-sales-collateral-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: Sales Collateral, Decks & Battlecards research operator
JOB: Build decks, one-pagers and battlecards from approved claims, current proof and versioned source material.
INPUTS: claim registry, product proof, competitor evidence, audience
EVIDENCE: approved product truth, proof reference, competitor source, audience brief
DETERMINISTIC: claim coverage, freshness, content diff
OUTPUT: deck outline, one-pager, battlecard
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/07-sales-collateral/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: Sales Collateral, Decks & Battlecards analyst operator
JOB: Build decks, one-pagers and battlecards from approved claims, current proof and versioned source material.
INPUTS: claim registry, product proof, competitor evidence, audience
EVIDENCE: approved product truth, proof reference, competitor source, audience brief
DETERMINISTIC: claim coverage, freshness, content diff
OUTPUT: deck outline, one-pager, battlecard
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/07-sales-collateral/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: Sales Collateral, Decks & Battlecards artifact operator
JOB: Build decks, one-pagers and battlecards from approved claims, current proof and versioned source material.
INPUTS: claim registry, product proof, competitor evidence, audience
EVIDENCE: approved product truth, proof reference, competitor source, audience brief
DETERMINISTIC: claim coverage, freshness, content diff
OUTPUT: deck outline, one-pager, battlecard
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/07-sales-collateral/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/07-sales-collateral/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# sales-collateral reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/07-sales-collateral/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/sales-collateral/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 162-184.

Pipeline Management & CRM Hygiene

System 08 · Pipeline Control Room. Make CRM state inspectable, preserve history and propose narrow updates instead of silently overwriting records. PDF pp. 185-207.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/pipeline-crm/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: 40fcd3923076c55c

Linked file: scripts/systems/08-pipeline-crm/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: phone cleared by omission

Linked file: skills/08-pipeline-crm-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: Pipeline Management & CRM Hygiene research operator
JOB: Make CRM state inspectable, preserve history and propose narrow updates instead of silently overwriting records.
INPUTS: CRM export, stage policy, activity receipts, owner map
EVIDENCE: current CRM read, field mapping, stage policy, activity receipt
DETERMINISTIC: field diff, merge candidates, stage consistency
OUTPUT: pipeline view, stale queue, field diff
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/08-pipeline-crm/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: Pipeline Management & CRM Hygiene analyst operator
JOB: Make CRM state inspectable, preserve history and propose narrow updates instead of silently overwriting records.
INPUTS: CRM export, stage policy, activity receipts, owner map
EVIDENCE: current CRM read, field mapping, stage policy, activity receipt
DETERMINISTIC: field diff, merge candidates, stage consistency
OUTPUT: pipeline view, stale queue, field diff
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/08-pipeline-crm/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: Pipeline Management & CRM Hygiene artifact operator
JOB: Make CRM state inspectable, preserve history and propose narrow updates instead of silently overwriting records.
INPUTS: CRM export, stage policy, activity receipts, owner map
EVIDENCE: current CRM read, field mapping, stage policy, activity receipt
DETERMINISTIC: field diff, merge candidates, stage consistency
OUTPUT: pipeline view, stale queue, field diff
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/08-pipeline-crm/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/08-pipeline-crm/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# pipeline-crm reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/08-pipeline-crm/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/pipeline-crm/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 185-207.

Deal Qualification & Forecasting

System 09 · Deal Risk & Forecast Desk. Separate stage labels, buyer-confirmed evidence, unresolved risk and probabilistic forecast assumptions. PDF pp. 208-230.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/qualification-forecasting/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: 4d79617e92a7d4d2

Linked file: scripts/systems/09-qualification-forecasting/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: verbal stage with no buyer next step

Linked file: skills/09-qualification-forecasting-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: Deal Qualification & Forecasting research operator
JOB: Separate stage labels, buyer-confirmed evidence, unresolved risk and probabilistic forecast assumptions.
INPUTS: opportunities, qualification policy, buyer evidence, amounts
EVIDENCE: buyer statement, CRM stage, amount source, next-step proof
DETERMINISTIC: criterion coverage, currency-safe totals, scenario ranges
OUTPUT: qualification packet, risk map, forecast ranges
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/09-qualification-forecasting/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: Deal Qualification & Forecasting analyst operator
JOB: Separate stage labels, buyer-confirmed evidence, unresolved risk and probabilistic forecast assumptions.
INPUTS: opportunities, qualification policy, buyer evidence, amounts
EVIDENCE: buyer statement, CRM stage, amount source, next-step proof
DETERMINISTIC: criterion coverage, currency-safe totals, scenario ranges
OUTPUT: qualification packet, risk map, forecast ranges
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/09-qualification-forecasting/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: Deal Qualification & Forecasting artifact operator
JOB: Separate stage labels, buyer-confirmed evidence, unresolved risk and probabilistic forecast assumptions.
INPUTS: opportunities, qualification policy, buyer evidence, amounts
EVIDENCE: buyer statement, CRM stage, amount source, next-step proof
DETERMINISTIC: criterion coverage, currency-safe totals, scenario ranges
OUTPUT: qualification packet, risk map, forecast ranges
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/09-qualification-forecasting/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/09-qualification-forecasting/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# qualification-forecasting reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/09-qualification-forecasting/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/qualification-forecasting/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 208-230.

Onboarding, Activation & Expansion

System 10 · Customer Growth Workspace. Track onboarding milestones, activation evidence, risk signals and expansion hypotheses without… PDF pp. 231-253.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/onboarding-expansion/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: 452c2ad009d44f66

Linked file: scripts/systems/10-onboarding-expansion/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: login mistaken for value

Linked file: skills/10-onboarding-expansion-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: Onboarding, Activation & Expansion research operator
JOB: Track onboarding milestones, activation evidence, risk signals and expansion hypotheses without…
INPUTS: customer plan, milestones, usage evidence, support context
EVIDENCE: signed scope, implementation plan, usage event, support context
DETERMINISTIC: milestone state, activation coverage, cohort math
OUTPUT: activation plan, milestone ledger, health report
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/10-onboarding-expansion/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: Onboarding, Activation & Expansion analyst operator
JOB: Track onboarding milestones, activation evidence, risk signals and expansion hypotheses without…
INPUTS: customer plan, milestones, usage evidence, support context
EVIDENCE: signed scope, implementation plan, usage event, support context
DETERMINISTIC: milestone state, activation coverage, cohort math
OUTPUT: activation plan, milestone ledger, health report
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/10-onboarding-expansion/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: Onboarding, Activation & Expansion artifact operator
JOB: Track onboarding milestones, activation evidence, risk signals and expansion hypotheses without…
INPUTS: customer plan, milestones, usage evidence, support context
EVIDENCE: signed scope, implementation plan, usage event, support context
DETERMINISTIC: milestone state, activation coverage, cohort math
OUTPUT: activation plan, milestone ledger, health report
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/10-onboarding-expansion/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/10-onboarding-expansion/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# onboarding-expansion reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/10-onboarding-expansion/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/onboarding-expansion/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 231-253.

Partnerships & Channel Enablement

System 11 · Partner Operations Hub. Prioritize partners, package enablement and track sourced opportunities without confusing attribution with causality. PDF pp. 254-276.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/partnerships/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: 397984f7fe0425c7

Linked file: scripts/systems/11-partnerships/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: revenue causal claim

Linked file: skills/11-partnerships-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: Partnerships & Channel Enablement research operator
JOB: Prioritize partners, package enablement and track sourced opportunities without confusing attribution with causality.
INPUTS: partner profiles, fit policy, enablement assets, opportunity records
EVIDENCE: signed agreement, territory rules, enablement completion, lead registration
DETERMINISTIC: partner tier, scope validation, duplicate lead ownership
OUTPUT: partner tier, enablement plan, co-sell packet
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/11-partnerships/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: Partnerships & Channel Enablement analyst operator
JOB: Prioritize partners, package enablement and track sourced opportunities without confusing attribution with causality.
INPUTS: partner profiles, fit policy, enablement assets, opportunity records
EVIDENCE: signed agreement, territory rules, enablement completion, lead registration
DETERMINISTIC: partner tier, scope validation, duplicate lead ownership
OUTPUT: partner tier, enablement plan, co-sell packet
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/11-partnerships/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: Partnerships & Channel Enablement artifact operator
JOB: Prioritize partners, package enablement and track sourced opportunities without confusing attribution with causality.
INPUTS: partner profiles, fit policy, enablement assets, opportunity records
EVIDENCE: signed agreement, territory rules, enablement completion, lead registration
DETERMINISTIC: partner tier, scope validation, duplicate lead ownership
OUTPUT: partner tier, enablement plan, co-sell packet
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/11-partnerships/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/11-partnerships/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# partnerships reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/11-partnerships/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/partnerships/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 254-276.

Claude Workflows & Sales Automation

System 12 · GTM Automation Orchestrator. Design Claude-assisted workflows with explicit state, tools, approvals, retries, receipts and recovery. PDF pp. 277-299.

Each dashboard view in this system follows the same contract:

Functional Dashboard

Operate the live state, queue and handoff - not a static summary.

Source-of-Truth Registry

Register authority, freshness, exact support and contradictions before reasoning.

Linked file: schemas/systems/sales-automation/source-packet.schema.json

Input & Data Workspace

Prepare evidence, authority, identity and missing fields before reasoning.

Identity & Entity Resolution

Resolve canonical IDs inside the correct tenant before scoring, merging, writing or reporting.

Canonical ID: 37a034d0c1b8d2bd

Linked file: scripts/systems/12-sales-automation/identity_resolver.py

Workflow Map

Carry source, restrictions and unresolved state through every step.

Research Workspace

Turn business questions into a bounded source plan, evidence map and explicit gaps.

Research Gap: prompt injection from source

Linked file: skills/12-sales-automation-evidence-research/

Research Prompt Compiler

Compile the source-builder prompt with exact authority, freshness, identity and output contract.

# Research Prompt Compiler
ROLE: Claude Workflows & Sales Automation research operator
JOB: Design Claude-assisted workflows with explicit state, tools, approvals, retries, receipts and recovery.
INPUTS: workflow goal, source systems, tool contracts, approval policy
EVIDENCE: workflow objective, tool contract, approval policy, state snapshot
DETERMINISTIC: graph validation, Skill/tool routing, idempotency key
OUTPUT: workflow graph, tool plan, approval matrix
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/12-sales-automation/research.md

Prompt Compiler

Compile named values into a source-bound prompt with an explicit output contract.

Analyst & Decision Prompt Compiler

Interpret deterministic results, compare alternatives and prepare a decision without side effects.

# Analyst & Decision Prompt Compiler
ROLE: Claude Workflows & Sales Automation analyst operator
JOB: Design Claude-assisted workflows with explicit state, tools, approvals, retries, receipts and recovery.
INPUTS: workflow goal, source systems, tool contracts, approval policy
EVIDENCE: workflow objective, tool contract, approval policy, state snapshot
DETERMINISTIC: graph validation, Skill/tool routing, idempotency key
OUTPUT: workflow graph, tool plan, approval matrix
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/12-sales-automation/analyst.md

Artifact Builder Prompt Compiler

Build a versioned, evidence-linked GTM artifact from approved inputs and decisions.

# Artifact Builder Prompt Compiler
ROLE: Claude Workflows & Sales Automation artifact operator
JOB: Design Claude-assisted workflows with explicit state, tools, approvals, retries, receipts and recovery.
INPUTS: workflow goal, source systems, tool contracts, approval policy
EVIDENCE: workflow objective, tool contract, approval policy, state snapshot
DETERMINISTIC: graph validation, Skill/tool routing, idempotency key
OUTPUT: workflow graph, tool plan, approval matrix
BOUNDARY: proposal != execution
RECOVERY: preserve valid state

Linked file: prompts/12-sales-automation/artifact.md

Tool & Execution Planner

Choose the minimum Skill, helper, server tool, connector or manual path and define the boundary.

Linked file: prompts/12-sales-automation/tool-planner.md

Skill Router & Anatomy

Package the repeatable procedure without turning access or execution into hidden assumptions.

Full Code Lab

Run exact calculations and state transformations outside the writing step.

Deterministic Code Lab B

Run state, handoff and reconciliation logic outside the writing step.

# sales-automation reconciliation
before = load_state()
changes = state_diff.run({"before": before, "after": candidate})
handoff = handoff_router.run(packet)
receipt = reconciler.run({
  "proposed_actions": proposed,
  "receipts": receipts
})
assert not receipt["ambiguous"]
save(handoff)

Linked file: scripts/systems/12-sales-automation/reconciler.py

Structured Output Schema

Validate packet shape, then review facts and business rules separately.

Spreadsheet & Model Lab

Use spreadsheet-style calculations with visible weights, units and missingness.

Integration & Automation

Separate triggers, Claude work, provider actions, approvals and reconciliation.

Approval, Action & Receipt Ledger

Keep proposed, approved, executed and reconciled states separate and inspect the receipt.

Linked file: schemas/systems/sales-automation/action-ledger.schema.json

QA & Review

Inspect evidence, restrictions, calculations and action claims before release.

Failure - Repair - Re-run

Preserve valid work and repair only the failed dependency.

KPI & Analytics

Keep units, periods, denominators and causal limits explicit.

Scenario A - Nominal

Run the normal case from input to reviewed output.

Implementation & Rollout

Move the system into a bounded pilot with files, tests and rollout evidence.

PDF pp. 277-299.

Closing

Build the operating system behind the revenue motion. Use Claude to preserve context, structure evidence, create the first usable artifact and surface the exact decision that still needs a human owner.

Primary Anthropic sources referenced

Sources checked 18 October 2026. Product capabilities, URLs, models and availability can change; recheck before implementation. The full method is in Chapter 02.

About this guide

NipPro AI — Specialized AI agents for sales, marketing and SEO. A NipPro AI publication by Alain Nono, Co-Founder & CEO. This training is offered to you by NipPro AI.

Independent guide. Not affiliated with or endorsed by Anthropic.

Sources checked 18 October 2026. Free to read on nipproai.com. A NipPro AI publication.