Engineering
Intelligent
Systems
dancodea.tech
Eliminating triage bottlenecks with AI-driven scoring
High-growth B2B sales teams frequently struggle with inbound lead triage. Valuable prospects sit idle in form submission databases while sales reps sift through spam, unqualified inquiries, and duplicate records.
We designed and deployed an autonomous, event-driven intake and qualification engine built on n8n, integrating Hunter.io for firmographic data enrichment and Google Gemini 1.5 for structured AI scoring.
Custom JavaScript normalizes email formatting, strips noise from international phone numbers, and performs instant lookups against the Master Database. Incomplete or email-less submissions are routed to an isolated quarantine sheet.
Returning prospects are compared field-by-field. If crucial intent signals change (e.g. upgraded job role, larger company, new project scope), the pipeline dynamically triggers re-enrichment and updated AI scoring.
Fetches verified employee headcount, primary industry vertical, and headquarters location in real time, creating an enriched profile prior to LLM evaluation.
A structured LangChain LLM chain powered by Gemini 1.5 evaluates authority, company fit, and explicit buying intent, outputting a strict JSON rating, qualification tier, and a 2-sentence rationale.
Highly Qualified prospects instantly trigger rich Slack Block notifications with direct 1-click email action buttons for reps, while automated cleanup keeps the central Master Sheet completely synchronized.
System Architecture & Pipeline Execution
Stage 1: Intake & Normalization
Webhooks capture form payload, standardize strings, and perform instant email deduplication checks.
Stage 2: Change Detection
Identifies existing contacts and verifies if significant attributes updated before triggering re-evaluation.
Stage 3: Hunter.io & Gemini AI Scoring
Enriches firmographics via API and executes deterministic LLM scoring across 4 qualification tiers.
Stage 4: Tier Routing & Slack Alerts
Dispatches leads to respective tier sheets and pings the sales team Slack channel with actionable CTAs.
Stage 5: Master Central Synchronization
Maintains data integrity by aggregating all tier records, removing obsolete copies, and syncing the Master Sheet.
Workflow Testing & Live QA Verification
Extensive live end-to-end test runs were performed to validate form intake, deduplication logic, Hunter.io firmographic enrichment, deterministic Gemini AI scoring, multi-tier Google Sheets synchronization, and instant Slack dispatch. Below are the verified live audit records and real-time team notifications.
1. Google Sheets Central Master Database
Comprehensive production intake sheet aggregating all incoming submissions enriched with Hunter.io company size, industry classification, geographic location, and verified domain metadata.
2. Highly Qualified Leads Tier
Automated segmentation routing top-tier prospects (AI Score 85+) with verified buying intent, budget alignment, and clear project implementation scopes.
3. Unqualified & Edge-Case Quarantine
Automated quarantine routing low-intent submissions, test entries, and non-business inquiries (AI Score 10–15) with detailed rationale.
4. Instant Slack High-Intent Lead Dispatch
Real-time Slack Block notification broadcasted to the sales channel featuring Gemini AI qualification breakdown, intent analysis, and a 1-click [Email Lead] action button for under 30s response time.
We build tailored autonomous AI systems.