- Artificial Intelligence
How Auriga IT Built an AI-Powered Search and Data Platform for VoomiSupply
How Auriga IT Built an AI-Powered Search and Data Platform for VoomiSupply
We replaced VoomiSupply's Searchanise plugin with a custom AI-powered search, data enrichment, and analytics platform. The system handles 1.5M+ HVAC products, understands buyer intent through semantic search, enriches product data in real time, and gives stakeholders live dashboards for every metric that matters.
01 / The Client
About VoomiSupply
A technology-driven HVAC and industrial supply marketplace built for the trades, serving contractors, technicians, and homeowners across the United States.
What VoomiSupply Does
VoomiSupply is a digital supply marketplace where buyers can purchase air conditioners, furnaces, mini-split systems, HVAC parts and tools, plumbing, and electrical components. Think of it as a specialized Amazon for technicians and home repair professionals, focused entirely on building systems and trade supplies.
Company Snapshot
Founded in 2019 and headquartered in Dresher, Pennsylvania, VoomiSupply is a technology-first B2B and retail e-commerce platform. It offers over 1.5 million products across HVAC, plumbing, electrical, and industrial supply categories, backed by AI-powered procurement and competitive pricing.
02 / The Challenge
What Was Broken
VoomiSupply was scaling fast, but its data infrastructure and product discovery tools could not keep up.
Unstructured Product Catalogue
Over 1.5 million products on Shopify with inconsistent, incomplete, and poorly classified data, making search and filtering unreliable for buyers.
No Data Enrichment Pipeline
Product specs like SEER, tonnage, BTU, and zones were missing or inaccurate. Daily price and inventory updates required manual effort with no automation.
Keyword Search Could Not Handle Intent
The existing Searchanise integration worked on keyword matching only. A buyer searching for "quiet bedroom AC" would get zero relevant results because no product literally contained those words.
No Visibility Into Business Metrics
Stakeholders had no centralized view of search analytics, product performance, stock distribution, conversion ratios, or traffic quality.
03 / What We Built
The Solution: Four Integrated Systems
We did not just fix search. We built an end-to-end intelligent data and discovery platform, replacing Searchanise with a purpose-built AI alternative.
AI-Powered Semantic Catalogue Search
A custom search engine that understands buyer intent, not just keywords. When a technician searches "3-ton Goodman heat pump," the system returns precise results even if the product title is structured differently. It maps natural language queries to the right product categories using semantic understanding.
Related search terms are automatically generated by AI and linked to relevant products, so the admin team does not have to manually tag every SKU.
Automated Search Admin Tooling
An AI-powered admin interface, Cygnus Copilot, where the VoomiSupply team can enrich and manage search behaviour using simple prompts. An admin can ask "How many products do we have for gas furnace?" and get instant, accurate answers. They can then link search terms, apply synonyms, and update product visibility, all driven by AI.
Cygnus Copilot is built on the CygnusAlpha platform with a three-layer memory architecture: situational awareness (active campaigns, stock alerts), corrective memory (supervisor-taught lessons), and reflective memory (self-learned patterns from past interactions). This means the system gets progressively smarter over time without manual prompt engineering.
Real-Time Data Ingestion and Enrichment Pipelines
Two automated pipelines run between VoomiSupply's source systems and the platform. The first handles semantic search indexing. The second handles product data enrichment, automatically classifying products by type, SEER rating, SEER2 rating, tonnage, BTU, zone configuration, and more. All pipelines run in real time, processing daily updates without human intervention.
Data Visualisation and Analytics Dashboard
An AI-powered analytics dashboard, Cygnus Insights, gives VoomiSupply stakeholders real-time visibility into product distribution, stock levels, pricing tiers, and technical specifications. Search-related insights include query volume, add-to-cart metrics, and conversion ratios. A traffic monitoring module also enables IP blocklisting for unwanted bot or scraper traffic.
Platform In Action
Inside the Dashboard and Copilot
Live screenshots from the Cygnus Insights dashboard and AI Copilot, the tools we built for VoomiSupply's team.
04 / Before vs. After
The Transformation
| Dimension | Before (Searchanise) | After (Auriga's AI Platform) |
|---|---|---|
| Product Search | Keyword-only matching, no intent understanding | Semantic search with AI query mapping and related terms |
| Data Quality | Unstructured, incomplete product data | AI-enriched with specs (SEER, BTU, tonnage, zones) |
| Admin Workflow | Manual tagging and search term management | Prompt-based AI admin: ask a question, get results |
| Data Pipelines | Manual updates, delayed syncing | Real-time automated ingestion and enrichment |
| Analytics | No centralized dashboard or insights | AI-powered dashboards with search, stock and traffic data |
| Traffic Protection | No visibility into bot or scraper traffic | IP blocklisting and traffic pattern monitoring |
05 / Key Outcomes
What Changed for VoomiSupply
The platform went from a basic Shopify storefront with keyword search to an intelligent, data-driven marketplace.
Search That Understands Buyers
Semantic AI search connects buyer intent to the right products, not just keyword matches.
Zero Manual Data Entry
Real-time pipelines handle all product enrichment, stock updates, and new listing ingestion automatically.
Data-Driven Decisions
Stakeholders now have live visibility into product performance, search trends, and conversion ratios.
AI Admin Copilot
The team manages catalogue queries through natural language prompts. No SQL, no spreadsheets.
Traffic Quality Control
Built-in traffic monitoring and IP blocklisting protects the platform from unwanted bots.
Self-Learning Agents
Built on a three-layer memory architecture, the platform learns from every interaction and gets progressively smarter without manual prompt updates.
06 / Technology
Key Capabilities Delivered
Semantic Search Engine
AI-powered product discovery with natural language understanding
Data Enrichment AI
Automated classification by product type, SEER, tonnage, BTU, zones
Real-Time Pipelines
Source-to-platform ingestion with intermediary enrichment
CygnusAlpha Platform
Three-layer memory architecture: situational, corrective, and reflective learning
LangGraph + Neo4j
Agent orchestration with lesson relationship graphs for progressive intelligence
PostgreSQL + pgvector
Vector-powered semantic retrieval with multi-tenant namespace isolation
07 / Frequently Asked Questions
Common Questions
VoomiSupply is a technology-driven B2B and retail e-commerce marketplace based in Pennsylvania, USA. Founded in 2019, it offers over 1.5 million products across HVAC, plumbing, electrical, and industrial supplies, serving contractors, technicians, and homeowners.
VoomiSupply had unstructured product data across 1.5M+ SKUs on Shopify. Their Searchanise search tool only supported keyword matching, product enrichment was manual, data pipelines did not exist, and there was no analytics dashboard for stakeholders.
We replaced Searchanise with a custom-built AI semantic search engine that understands buyer intent and maps natural-language queries to correct products. It also auto-generates related search terms using AI.
Cygnus Copilot is a conversational AI assistant built for VoomiSupply's admin team. It allows non-technical stakeholders to query the product catalogue using natural language prompts and get instant, accurate answers.
We built two real-time automated pipelines. The first handles semantic search indexing. The second manages product data enrichment, classifying products by SEER ratings, SEER2, tonnage, BTU, zones, and product type.
Searchanise operates primarily on keyword matching. VoomiSupply needed semantic understanding, AI-powered data enrichment, custom admin tooling, and integrated analytics, capabilities that required a purpose-built solution.
Cygnus Copilot runs on the CygnusAlpha platform, built with LangGraph for agent orchestration, PostgreSQL with pgvector for semantic retrieval, and Neo4j for lesson relationship tracking. It uses a three-layer memory architecture (situational, corrective, reflective) that allows agents to learn and improve from every interaction without manual prompt engineering.
Auriga IT is a digital solutions company founded in 2010 by IIT Roorkee alumni. Based in Jaipur, India with offices in Canada and Gurugram, we have delivered 1,000+ projects across AI, data engineering, e-commerce, and enterprise software.
08 / Related Work
More AI Case Studies from Auriga IT
Explore how we have applied AI, data engineering, and intelligent automation across industries.
AI-Powered Omnichannel Support for Ferns N Petals
One AI-powered inbox unifying multi-channel queries for faster, smarter, seamless customer support across a major e-commerce brand.
Natural Language Querying for Relational Databases
Enabling non-technical users to query complex relational databases using plain English, powered by AI and NLP.
Secure Self-Hosted AI Assistant for CRM
A privacy-first AI assistant deployed on-premise to help sales teams interact with CRM data through conversational queries.
Intelligent Knowledge Base with Generative AI
Building a generative AI-powered knowledge base that understands context and delivers accurate answers from enterprise documentation.
Browse all AI case studies at Auriga IT / Our Work / AI
Ready to Build Something Like This?
Let's talk about your e-commerce data challenge
Whether it is search intelligence, data pipelines, or AI-powered analytics, Auriga IT builds it from the ground up.
Talk to Auriga ITRelated content
Auriga: Leveling Up for Enterprise Growth!
Auriga’s journey began in 2010 crafting products for India’s [...]
Stay Close to What We’re Building
Get insights on product engineering, AI, and real-world technology decisions shaping modern businesses.
