Product marketing · Positioning
positioning
Positioning Intelligence Loop
trigger: weekly cron + every closed-won / closed-lost deal
Fathom / FirefliesCall miningWin/loss transcripts, objections, exact buyer language
→
PythonCompetitor scrapeFirecrawl pulls review sites + competitor messaging
→
Claude CodePattern analysisObjection clusters, win themes, whitespace vs competitors
→
ClaudePositioning vNUpdated narrative, messaging hierarchy, proof points
→
n8nPropagationNew copy pushed to sequences, landing pages, decks
→ 4 positioning shifts shipped in the current build - each one validated in-market before the next
the loop that keeps messaging honest: buyers write the positioning, the system just listens
Copywriting
copywriting
7-Stage Copy Engine
trigger: new enriched account lands in the Clay table
Clay + LLMRole mappingPersona classified against ICP definition
→
Ahrefs APISEO intelRankings, traffic, current SEMrush/Ahrefs stack usage
→
LLMGap analysisWhere their current stack underperforms, with evidence
→
LLMStrategic anglePositioning picked per account, mapped to pain
→
ClaudeEmail generationPer-contact copy in approved voice + QA pass
→
SmartleadShipStraight to sequence - zero human review
→ Hyper-personalized copy at scale, 1.5%+ positive replies across 150+ domains
the stage most teams skip is 3 and 4 - that is exactly where generic AI copy dies
Outbound
outbound
Signal-Routed Outbound
trigger: job posting, tech-stack change, or LinkedIn engagement detected
TheirStack / TrigifySignal captureHiring patterns, stack changes, post engagers
→
n8nSignal routerBranch by signal type, strength, and recency
→
ClayWaterfall enrich5 sources: 85%+ coverage, +45% accuracy
→
LLMSequence pickMessage angle matched to the triggering signal
→
Smartlead + HeyReachMulti-channelEmail + LinkedIn, 300+ mailboxes, 95%+ deliverability
→ 40% pipeline-efficiency gain, $180K+ ARR growth in 6 months
fit: TAM above ~20k accounts, LTV above $5K - below that, run ABM instead
ABM
abm
Asset-Led ABM Engine
trigger: account enters tier-1 target list
PythonDeep researchPlaywright + asyncio fan-out, 100+ domains per batch
→
Claude + OpenAIContext-to-contentPer-account narrative from research + Ahrefs data
→
Pages + SendsparkPersonal assets500+ visit-tracked landing pages, per-account video
→
TrackingEngagement watchShare of Answer + AI Share of Voice reported per page
→
HubSpotSales triggerVisit + watch signals alert the closer with context
→ Enterprise logo closed at $25K MRR
fit: TAM under ~20k accounts, LTV above $50K - the assets do the room-warming
RevOps · Inbound
revops
Inbound Scoring & Routing
trigger: form fill, demo request, or RB2B-identified visitor
RB2B / FormsCaptureAnonymous visitors resolved to person + account
→
n8nWebhook intakeCo-designed schema with Product + Engineering
→
LLM rubricScoring agentCustom ICP rubric scores fit + intent, with reasoning
→
HubSpotAuto-routingScore + context pushed to properties, owner assigned
→
SlackHot-lead alertHigh scores ping the closer in minutes, not days
→ Lead qualification cut from 48h to 9.6h (-80%)
speed-to-lead is the cheapest conversion lift in B2B - this buys it permanently
LinkedIn content
content
Content Signal Flywheel
trigger: weekly content cycle + every post published
Transcripts + commentsPain miningReal ICP language harvested from calls and threads
→
ClaudeDraft in voiceFounder-voice posts from mined pains + proof points
→
TrigifyEngager captureLikers + commenters on target posts become records
→
ClayICP filter + enrichOnly in-ICP engagers pass, fully enriched
→
HeyReach + SmartleadWarm sequenceEngagement-referencing outreach, synced to CRM
→ Content stops being a vanity channel: every post becomes a 2P signal source feeding outbound
fit: any B2B motion where the ICP is active on LinkedIn