Sales & Revenue Templates
Four templates covering the revenue engine: lead handling, RFP response, quoting, and campaign operations.
create-leafmesh my-project --template sales-orchestration # or:
create-leafmesh my-project --template bd-proposals
create-leafmesh my-project --template quote-to-cash
create-leafmesh my-project --template marketing-opssales-orchestration — Lead → Handoff
12 agents · 3 scheduled crons
| Agent | Type | Role |
|---|---|---|
lead_intake_agent | LLM | Intake owner — normalises any-source lead (webform, demo request, event, partner), then enriches with firmographics, funding, tech stack, decision-makers |
lead_qualification_agent | LLM | Qualification rep — scores ICP fit + intent → cold/warm/hot/spam tier, then drafts the personalised opener citing specific signals |
sequence_orchestrator_agent | Programmatic | Tier × intent matrix → AE direct / SDR review / nurture |
deal_nurturer_agent | Programmatic | Multi-touch drip ledger; daily sweep; graduates to AE on engagement threshold |
abm_orchestrator_agent | Programmatic | 3+ contacts from one domain in 14 days → coordinated ABM mode, one AE owner |
sales_research_agent | LLM | Sales researcher — per-deal competitor intel pack or pre-meeting account dossier in one role |
lead_revival_agent | LLM | Weekly cron — re-checks 90-day-cold leads for job moves and funding events |
pipeline_health_agent | Programmatic | Pipeline health analyst — hourly SLA/queue sweep, opportunity-stage ledger, idle-deal warnings, per-rep coverage |
sdr_human / ae_human / sales_manager_human | Human | SDR review → AE ownership → manager escalation |
sales_audit_logger_agent | External | Immutable log of every routing, sequence, and handoff |
How it works. A lead lands → intake normalises and adds company context (registering the contact with ABM in parallel) → qualification tiers it and writes the opener → the orchestrator routes: hot to AE, warm to SDR review, cold to the nurture ledger. The nurturer drips touches daily and graduates leads to an AE when engagement crosses threshold. Cold leads aren't dead — the revival cron resurrects them on new signals.
Value. The expensive failure in sales is leads dying in the gaps — unworked queues, cold leads nobody re-checks, five SDRs hitting one account simultaneously. This mesh closes all three with hourly SLA enforcement, weekly revival sweeps, and account-level ABM dedup.
bd-proposals — RFP → Win
12 agents · 1 cron
| Agent | Type | Role |
|---|---|---|
rfp_intake_agent | LLM | Opportunity intake owner — scans RFP feeds on the daily cron, normalises inbound RFPs, and extracts structured requirements from attachments |
bid_no_bid_decisioner_agent | LLM | 5-axis fit scoring → bid / qualify / no-bid |
precedent_research_agent | LLM | Precedent researcher — pulls similar prior wins and keyword-searches the proposal corpus as drafting context |
section_drafter_agent | LLM | Drafts each required section |
pricing_configurator_agent | Programmatic | Pricing within bands; flags outliers |
contracts_review_agent | LLM | Contracts reviewer — regulatory/legal fit check plus clause-by-clause playbook review; unacceptable clauses → legal hard gate |
competitive_intel_agent | LLM | Competitive intelligence analyst — win/loss post-mortems plus the competitor-signal ledger, queryable by competitor and horizon |
proposal_writer / practice_lead / legal / partner | Human | Default-deny review chain by stakes |
bd_audit_logger_agent | External | Immutable audit |
How it works. Two front doors — RFPs arrive (email, portal) or get found (the intake cron). Either way: intake structures requirements → bid/no-bid scoring → precedent research feeds the section drafter → pricing → the dual-mode contracts review gate → the human chain signs by stakes. Lost deals feed the competitive-intel analyst, which feeds the next bid decision.
Value. Proposal teams burn senior hours rewriting boilerplate and miss RFPs they never saw. This mesh finds opportunities proactively, drafts from your own winning language, and hard-gates the clauses that create post-signature pain.
quote-to-cash — RFQ → Signed Quote
14 agents · 1 cron
| Agent | Type | Role |
|---|---|---|
rfq_ingester_agent | LLM | Normalises the RFQ |
cost_calculator_agent | Programmatic | COGS per line item |
discount_policy_agent | LLM | Tier matrix per customer + competitive signals |
approval_router_agent | Programmatic | Deterministic amount × discount matrix → rep / deal desk / finance / CFO |
deal_compliance_agent | LLM | Deal-compliance analyst — playbook clause check (auto-renewal, escalators, SLA, liability, data residency, IP) plus ASC 606 rev-rec classification; unacceptable clauses → legal hard gate, rev-rec ambiguity → CFO |
quote_drafter_agent | LLM | Assembles the final quote |
bundle_recommender_agent | LLM | Cross-sell: adjacent SKUs from similar accounts |
contract_redline_tracker_agent | Programmatic | Redline ledger; stalled-redline reports |
competitor_watch_agent | Programmatic | Daily cron — competitor price changes refresh discount tiers |
sales_rep / deal_desk / finance_approver / cfo | Human | Four-tier default-deny approval ladder |
qtc_audit_logger_agent | External | Immutable audit |
How it works. RFQ → COGS → discount tier → the deterministic router (no LLM decides who approves money). The deal-compliance analyst playbook-checks contract clauses before the CFO sees the deal and classifies revenue recognition for finance in the same role. Redlines are tracked as first-class state so deals don't silently stall.
Value. Quote desks leak margin two ways: discounts that creep past policy, and non-standard clauses signed in a hurry. Here the approval matrix is code, the clause check is systematic, and daily competitor pricing keeps tiers honest.
marketing-ops — Campaign & Growth
15 agents · 2 crons
| Agent | Type | Role |
|---|---|---|
campaign_strategist_agent | LLM | Campaign strategist — normalises the incoming brief, then sets objectives, KPIs, channel mix, budget allocation |
market_intelligence_agent | LLM | Market intelligence analyst — audience segments + persona scoring plus the competitor landscape for the brief |
copywriter_agent | LLM | Three variants per channel (A/B/C test-and-learn) |
visual_briefer_agent | LLM | Creative brief for image generation or designer |
compliance_reviewer_agent | LLM | Brand-safety + legal-claim gate before approval |
social_media_agent | LLM | Per-platform variants with char limits enforced; reply triage → crisis path on anger |
publisher_agent | Programmatic | Multi-channel publish; every impression tagged for attribution |
marketing_analytics_agent | LLM | Marketing analytics lead — KPI-vs-target judgment, multi-touch revenue attribution, email engagement vs segment baselines |
brand_guardian_agent | LLM | Brand guardian — tracks mention sentiment, detects spikes, and drafts the fast-path crisis response to CMO and legal |
aeo_geo_optimiser_agent | Programmatic | Weekly cron — scores pages for AI answer-engine pickup (FAQ structure, citations, schema) |
campaign_orchestrator_agent | Programmatic | Weekly sweep across live campaigns |
marketing_manager / legal_reviewer / cmo | Human | Approval, regulated claims, crisis/budget tiers |
marketing_audit_logger_agent | Programmatic | Immutable campaign audit |
How it works. The strategist normalises the brief and pulls market intelligence — audience and competitor in one analyst — then sets the strategy → copy and visuals → compliance gate → human approval → publish with attribution tagging. After launch the loop continues: the analytics lead judges performance and attribution → the next brief. Two sentinels run continuously: the brand guardian (sentiment spike → its own crisis fast-path) and the AEO auditor — keeping content visible to AI answer engines, not just search.
Value. Marketing's hidden tax is coordination — briefs lost in chat, claims that should have seen legal, campaigns nobody measured. The mesh makes the pipeline explicit and gated, and the AEO loop addresses the channel shift nobody's staffed for: buyers asking AI instead of searching.