Knowledge Graph Platform / Graph analytics · AI · Explainable intelligence

Turn fragmented data into queryable, explainable & actionable intelligence.

A knowledge graph platform for the modern enterprise.

GraphEdge Analytics transforms unstructured and fragmented data into explainable, queryable & actionable intelligence using graph analytics and AI. Stop chasing dashboards. Start asking your business a question and getting an answer it can defend.

00 / Why GraphEdge

From scattered data to connected decisions.

The knowledge is there.
The understanding isn't.

Every organization runs on knowledge that is scattered across unstructured sources — documents, filings, emails, contracts, reports, logs — and disconnected from the structured data that gives it meaning. The facts that matter exist, but they're fragmented, unlabeled, and locked in systems that were never built to talk to each other. Search returns documents. Dashboards return numbers. Neither returns understanding — so emerging risks, hidden exposure, and real opportunities stay buried until they surface the hard way.

One connected ecosystem of knowledge.

GraphEdge turns that fragmented data into an actionable ecosystem of knowledge. We gather the relevant data, model its ontology, extract the relationships between entities, and marry in your structured context to build a connected knowledge graph. Then we apply AI engineering, NLP, and data-science techniques so your team can understand it in plain language and act on it — and every insight keeps its evidence trail, pointing back to the exact source documents behind it.

01Gather 02Model the ontology 03Build the relationships 04Marry in structured data 05Engineer the knowledge ecosystem 06Keep the evidence trail 07Understand it with NLP 08Act on it with data science

A decision layer, not just a map.

The result isn't another diagram of what's connected — it's a decision layer over fragmented business knowledge. Teams move from discovery to action: spotting newly emerging risks, comparing exposure across companies and suppliers, tracing the evidence behind every answer, and knowing what to investigate next. Decisions become faster, defensible, and grounded in the full picture instead of a fraction of it.

Deploys on
AWS · Azure · Google Cloud · On-prem · Your VPC, your choice
01 / The Products

One engine. Products that ship.

GraphEdge isn't a slide deck. The same engine already powers two verticals — Silent Facts, live over the public markets, and FairMarket IQ, live and a work in progress for public procurement — each turning fragmented, unstructured data into an explainable, queryable knowledge graph.

Live in production PRODUCT / 01

Silent Facts

https:// silentfacts.com live

Institutional-grade financial intelligence over the public markets.

01 — What we solve

Analysts and risk teams drown in public data — thousands of SEC filings across hundreds of companies, plus a parallel universe of federal contracts and awards. The relationships that actually signal risk and opportunity (shared suppliers, board overlaps, hidden exposure, government dependence) stay buried in unstructured text and disconnected from each other.

02 — How we solve it

Silent Facts ingests every filing across 750+ public companies and federal spending data from USAspending.gov — contracts, awards, and obligations — then extracts the entities and the relationships between them and connects everything into a single knowledge graph. Ask in plain English and get a typed, navigable answer — every edge traced back to the exact filing or federal award behind it.

03 — Why it matters

Weeks of manual document review collapse into seconds. Surface supply-chain exposure, board interlocks, emerging risk, and federal-revenue dependence across the whole market — with a defensible evidence trail under every answer. FDA data is next — drug approvals, trials, and enforcement actions joined into the same graph, extending the same relationship intelligence into life sciences.

Visit silentfacts.com
Live · Work in progress PRODUCT / 02

FairMarket IQ

https:// fairmarket-iq.com WIP

Public-procurement intelligence — equity matchmaking & compliance.

01 — What we solve

Public agencies require prime contractors to award a real share of every project to certified small, minority, women, and veteran-owned businesses. Primes struggle to find capable, compliant subcontractors in the right trade and region — and agencies can't easily spot "front companies" or pay-to-play conflicts hidden across disconnected vendor directories.

02 — How we solve it

FairMarket IQ unifies public award listings, supplier and certification registries, and debarment and campaign-finance records into one connected knowledge graph. It matches newly awarded primes with vetted diverse subcontractors, then instantly audits the hidden relationships — shared officers, shared addresses, debarment status — that signal fraud risk.

03 — Why it matters

Primes hit their diversity targets with zero front-company or debarment liability. Certified subcontractors get high-intent leads the moment a relevant contract is awarded. Agencies catch collusion and pay-to-play violations before contracts are ever signed.

Visit fairmarket-iq.com
Real engine output Screenshots from Silent Facts, running on the GraphEdge engine
ask> Show AMD's customers and their suppliers across the semiconductor industry
GraphEdge knowledge graph of AMD's customers and their suppliers across the semiconductor industry — AMD, FTNT, AVGO, INTC, MU and MRVL plus customers including MSFT, Sony and OpenAI, with every edge typed 'SUPPLIES' and an AI-written explanation generated from the graph.
SUPPLY-CHAIN INTELLIGENCE · LIVE ON SILENTFACTS.COM

A plain-English question becomes a typed, navigable supply-chain graph — every "SUPPLIES" edge traced to its source filing, with the written explanation generated from the graph itself.

ask> Show AMD's board interlocks — the directors it shares with other public companies
GraphEdge knowledge graph of AMD's board interlocks — the directors AMD shares with other public companies, each shared-director edge typed and traced back to its source filing.
BOARD INTERLOCK · AMD

One question surfaces every director AMD shares with another public company — the hidden board-level network that signals influence, conflicts, and concentration risk. Every shared-director edge is typed and cited to its filing.

02 / Why Graphs

Your data is connected. Your insights should be, too.

Most enterprises sit on millions of unstructured documents, emails, tickets, contracts, transactions, and logs — scattered across warehouses, lakes, SaaS apps, and shared drives. Vector search retrieves chunks of text. It loses the relationships. GraphEdge captures the relationships between every entity — people, processes, products, customers, contracts — and gives AI the context it needs to be genuinely useful.

80%+
of enterprise data is unstructured
documents · emails · tickets · transcripts · logs
12+
disconnected systems per Fortune 500
CRM · ERP · HRIS · ticketing · DMS · email · data lake
1
unified, queryable knowledge graph
every entity · every relationship · every source cited
DOCUMENT RETRIEVAL

Vector Search

Disconnected text chunks. Context lost at retrieval.

  • No relationships preserved between sources
  • Hallucinations and shallow synthesis
  • Provenance lost at the chunk boundary
KNOWLEDGE GRAPH

GraphEdge

Customers, contracts, products, risks — typed and linked.

  • Relationships are first-class context
  • Every answer cites the source record
  • Synthesis across the entire enterprise
03 / The Platform

Three capabilities. One graph-native engine.

GraphEdge Analytics is a domain-agnostic platform. Connect your structured, semi-structured, and unstructured data — the engine builds a queryable knowledge graph and exposes it through plain English, with full provenance.

01

Graph-native retrieval

We don't just retrieve text — we traverse relationships to find deep, synthesis-level answers across your entire enterprise.

retrieval: graph-native
02

Explainable AI

Stop trusting black boxes. Every AI-generated insight links back to the source record, paragraph, and graph path that produced it.

provenance: ✓ every answer
03

Enterprise-ready scalability

Production-grade architecture that handles growth from pilot to global rollout. Deploy on any cloud — AWS, Azure, GCP — on-prem, or inside your own VPC.

deploy: any cloud · VPC · on-prem
04 / For Corporates

Where graph-native intelligence changes the game.

The same engine, applied to the questions that move the needle for enterprise leaders — finance, operations, customer, and risk.

FINANCE CASE / 01

Financial Integrity

Detect fraud rings by analyzing transactional flows, not just single transactions. Surface money-laundering, kickbacks, and conflict-of-interest patterns hidden in entity relationships.

finds multi-hop fraud rings
OPERATIONS CASE / 02

Operational Intelligence

Map your entire supply chain — vendors, parts, logistics, contracts — to identify bottlenecks and concentration risk before they hit the P&L.

predicts disruptions early
CUSTOMER CASE / 03

Customer 360

Link every interaction across CRM, support, billing, and product into a single graph. Deliver hyper-personalized experiences and spot churn before it happens.

unifies every touchpoint
LEGAL & COMPLIANCE CASE / 04

Contract & Obligation Intelligence

Every clause, party, deadline, and dependency — connected. Answer "which contracts expose us to X?" in seconds instead of weeks of manual review.

tracks obligations & risk
RISK CASE / 05

Enterprise Risk Mapping

Concentration, counterparty, regulatory, and cyber risk — modeled as a single typed graph. Run "what-if" traversals across the dependency network.

models interconnected risk
KNOWLEDGE CASE / 06

Internal Knowledge Search

Wikis, decks, tickets, emails, and transcripts unified into one graph. Employees ask plain-English questions and get cited, traceable answers.

answers with provenance
05 / Why Now

Three curves just crossed.

Knowledge graphs aren't new — but the cost of building one over a live enterprise data set collapsed in the last 24 months. Three independent curves crossed at the same time, and graph-native AI went from research project to production-ready.

01

LLMs cleared natural-language graph queries

Plain-English questions now compile to graph traversals at >90% accuracy on the shapes business users actually ask.

60%  →  90%+
02

Graph infrastructure got cheap

What used to be six-figure enterprise contracts is now consumption-priced. The cost floor for a live, queryable graph collapsed.

$100K/yr  →  <$500/mo
03

Enterprise data went machine-first

SaaS APIs, data lakes, and document stores ship structured JSON now. Ready for ingestion at scale — no more screen-scraping or brittle ETL.

PDF/HTML  →  JSON

Three years ago this stack cost $300K/yr. Today: $4K/mo. That's the unlock.

06 / How it works

From scattered data to graph-native answers.

  1. 01 Ingest

    Connect your existing data sources.

    SQL, NoSQL, APIs, data lakes, document stores, email, chat, ticketing — structured or unstructured. We meet your data where it already lives.

    SQLNoSQLS3SharePointSalesforceJiraREST APIs
  2. 02 Model

    The engine builds your knowledge graph.

    LLM-driven entity and relationship extraction. Identity resolution across systems. Schema inference tuned to your business — not a generic ontology.

    Entity ResolutionEdge ClassificationSchema GroundingGraph Embeddings
  3. 03 Ask

    Question your business in plain English.

    Plain-English questions return answers backed by the exact graph paths that produced them. Synthesis across the whole enterprise — with full provenance.

    Natural languageGraph traversalProvenance trailsSub-second p95
07 / The Team

Built by a founder who ships.

GraphEdge Analytics is a founder-led company. We've already taken the engine from idea to a product live in production — Silent Facts — and we're building in public toward enterprise pilots.

Saikat Maitra, Founder of GraphEdge Analytics

Saikat Maitra

Founder & CEO

Saikat Maitra is the Founder and CEO of GraphEdge Analytics. He brings 20+ years of experience across data, enterprise software, cloud platforms, analytics, and business application delivery, with industry experience spanning shipping, manufacturing, investment banking, and healthcare technology. Most recently, he served as an Enterprise Architect at Johnson & Johnson. His work has focused on enterprise data platforms, cloud architecture, analytics, integration, and turning complex data environments into usable business applications.

After two decades working with structured and relational data, Saikat saw that much of an organization's most valuable knowledge remains trapped inside documents, filings, reports, contracts, policies, and other unstructured sources.

He founded GraphEdge Analytics to turn scattered enterprise information into connected, explainable, and queryable intelligence using AI engineering, knowledge graphs, retrieval-augmented generation, natural language processing, and data science.

Saikat is currently leading the development of the GraphEdge platform and its first product, Silent Facts — a live SEC intelligence product that maps public-company filings into a knowledge graph across 750+ companies in the S&P 500, Nasdaq, and Dow.

Knowledge graphs Applied AI Data engineering 0→1 product
08 / Get in touch

Let's map your knowledge graph.

Pilots on enterprise data. Demos of the platform. Conversations with anyone serious about graph-native intelligence.

STATUS Building in public · Pilots open now
TO smaitra1@graphedgeanalytics.com

We'll reply within 1 business day. No spam, ever.