In DevelopmentSaaS

AI Storm Expansion Engine

Storm intelligence and campaign activation for home service companies — see who was hit and act.

What it solves

Storm-driven home service companies lose money because they cannot quickly tell which leads and past customers were affected by a storm, which neighborhoods deserve door knocking, or which reps to deploy where. Most rely on scattered tools, manual maps, spreadsheets, screenshots, and guesswork. The result is slow, disorganized follow-up after weather events.

Built for Storm-driven home service companies — roofing, restoration, exterior remodeling, solar, gutter, siding, and window companies — plus their sales, canvassing, and marketing managers.

Overview

The app turns severe storm events into organized sales opportunities for home service companies. Users import CRM leads and past jobs, view storm events and property pins on a map, and set severity trigger rules that flag affected properties. A matching engine scores each property by storm severity and lead value, and users build campaign lists, assign reps or crews, and export outreach. An AI layer recommends the best campaign type and messaging after each event, and an owner dashboard reports business impact. It is built as a multi-tenant SaaS with role-based access.

What this demonstrates

It implements geospatial matching of storm geometry against geocoded properties with radius and severity scoring, plus lead-value and priority scoring. It shows a provider-adapter layer that abstracts storm data sources with encrypted API keys, multi-tenant data isolation with role-based permissions, CRM import with deduplication, and a grounded AI recommendation layer that only uses verified system data.

What makes it different

This is not just a storm map. This is a revenue activation engine — a CRM-connected storm response system that turns weather events into organized sales opportunities.

Answer one question fastWho was hit, how badly, and what should we do next?
Turn storms into revenueConvert weather events into actionable, organized outreach instead of guesswork.
Understandable in 30 secondsThe core flow is simple enough for a first-time user to grasp immediately.
Replace scattered toolsOne workflow instead of manual maps, spreadsheets, screenshots, and cross-referencing.
Multi-tenant and permission-scopedEvery record belongs to an organization; no cross-tenant access.

Core capabilities · 14

Storm map dashboard
Interactive map showing recent storm events, paths, hail/wind zones, and property pins.
Storm provider integration layerProvider adapter
Adapter system that ingests storm events and geometry from one or more data providers.
CRM lead import
Upload and field-match CRM leads, with manual lead creation as a fallback.
Property geocoding
Auto-geocodes addresses so every property has coordinates before storm matching.
Past job tagging
Labels completed jobs and past customers for warranty and follow-up targeting.
Storm trigger rules
Plain-language rules that flag properties by severity thresholds and radius.
Affected property detection engineDetection engine
Geospatial matching that determines which properties each storm event affected.
Severity and priority scoring
Scores affected properties by storm severity, lead value, and priority.
Campaign list builder
Turns affected properties into lists for mail, SMS, email, calls, or door knocking.
CSV export
One-click export that excludes do-not-contact leads and only permitted fields.
Rep and crew assignment
Assigns affected properties to reps with territory grouping and a rep task view.
AI campaign recommendations
Suggests campaign type, audience, timing, and messaging using verified system data only.
CRM integration system
Modular import/export and webhook sync back to the source CRM.
Business impact dashboard
Tracks affected leads, campaigns, assignment progress, and estimated pipeline value.

How it works

Open the Storm Dashboard
See recent storm events on the map
View CRM leads and past jobs as property pins
Select a storm event or storm path
System identifies affected properties
System scores each property by severity and lead value
Generate a campaign list
Export, assign, or launch the campaign

How the score works

Priority Score

Severity Score
From hail size, wind speed, tornado risk, storm path proximity, storm type, and event recency.
Lead Value Score
From lead status, past-customer status, estimated deal value, prior job value, rep status, and contactability.
Priority Score
Combination of severity, lead value, recency, and campaign readiness.
Match confidence
Confidence rating on whether the property was truly affected by the event.

Key screens

Storm DashboardDefault map view of the most recent severe storm activity with sync status.
LeadsLead list and detail views with status, geocoding, and past-job labels.
Affected properties tableFilterable table by priority, storm type, lead status, and campaign readiness.
CampaignsCampaign list creation, naming, export, and status tracking.
AssignmentsRep deployment, territory grouping, and progress by rep.
SettingsStorm provider connection, trigger rules, and account configuration.

Data model

Organization
User
WorkspaceSettings
Property
Lead
JobHistory
StormProviderConnection
StormEvent
StormSyncLog
StormTriggerRule
AffectedProperty
CampaignList
CampaignListItem
Assignment
CRMConnection
CRMSyncLog
AuditLog

How it is built

A Next.js and React application backed by Prisma and PostgreSQL (with PostGIS, provisioned via Docker Compose), rendering an interactive MapLibre map with clustered property pins. Storm matching and scoring run server-side, integration secrets are encrypted, and Claude powers optional AI recommendations with a deterministic fallback.

APIs and services

Anthropic
MapLibre

Classification

Industries: Home Services
Capabilities: Geospatial Analysis CRM AI Automation Multi-Tenant Architecture Reporting
Technologies: Docker Next.js Node.js React Tailwind CSS TypeScript PostgreSQL Prisma Zod

Building something with similar Geospatial Analysis, CRM, AI Automation functionality?

Start a project and reference this one — the conversation begins with what you have already seen.

Start a project