One invention — the gBAG — powers three live products across cybersecurity, regulatory compliance, and planning intelligence. Deterministic. Auditable. No LLMs in the decision path.
320 frameworks, 146K+ obligations. Full planning hierarchy (LEP, SEPP, DCP). Growing weekly.
Australian cadastre, zoning, heritage, contamination, and transport overlays. Starting from NSW, expanding nationally then globally.
4.5M business profiles. Industry classification, regulatory exposure, domain security posture.
"Every regulation, every network, every supply chain is a graph. Risk doesn't sit in one node. It flows through chains. Detecting risky chains before they break is the real Return on Intelligence."
We built an engine that reads flat documents, unstructured data, and complex rule systems and turns them into scored, navigable intelligence networks with Bayesian reasoning on every edge.
The gBAG — Geospatial Bayesian Acyclic Graph — applied to three domains. Each built in weeks from a shared architecture, not months from scratch. All live with self-serve sign-up.
Enterprise Security Intelligence
Scans your external attack surface, maps vulnerabilities into multi-step attack chains, and predicts which threats will be exploited next. Zero infrastructure access required.
Regulatory Compliance Intelligence
Upload a compliance document. Get it mapped against every regulatory obligation, scored across five risk dimensions with cascade probability analysis. Minutes, not months.
Planning Intelligence
Full planning graph. Enter any address, get every applicable control identified, conflicts resolved, and approval likelihood scored. Seven persona-specific report types in ~3 seconds.
No language model makes decisions. The graph reasons. The scores compute. LLMs only handle document ingestion and communication. That means no hallucinations, no black boxes, and full auditability.
Directed Acyclic Graphs with Bayesian reasoning layers on every edge. Risk cascades computed deterministically.
Gradient Boosted Trees for classification and scoring. Deterministic, auditable, under 50ms per prediction. No LLMs at the decision layer.
Every score traces back to its source. No hallucinations. No black boxes. The graph reasons, the AI communicates.
Exploit prediction (industry: 0.500)
Across 32.6M graph edges
Across 320 regulatory frameworks
Cascade probability tables
Our platforms run on enterprise-grade infrastructure with security, reliability, and compliance built into every layer — not bolted on after the fact.
A Geospatial Bayesian Acyclic Graph — a proprietary reasoning architecture that models rules, obligations, and constraints as scored probabilistic networks. Every node carries a compliance probability; risk cascades through dependencies. Point it at a new domain and you get a new product, a new market.
Cybersecurity Intelligence
507M+ threat records, 184 ML models, 422K+ attack chains. Exploit prediction at 0.837 AUC vs 0.500 baseline.
auditroi.com →Regulatory Intelligence
320 frameworks, 146K+ obligations, 12,278 rules scored, 1,732 enforcement actions. 26 jurisdictions, 25 industries.
auditdss.com →Planning Intelligence
Full planning graph. 8,130+ obligations, 613 DAG nodes, 7 persona-specific reports. Address to report in ~3 sec.
zonedss.com →All live with self-serve sign-up and Stripe billing. No LLMs in the decision path. The architecture extends to any graph-shaped risk domain — every new domain is a new graph, a new product, a new market.
We deploy the proven gBAG engine inside your organisation — with your data, your branding, and your workflows. Deterministic AI chat agents, regulatory assessments, report generators, geospatial intelligence, and API integrations.
Explore ServicesYour brand, your domain
Dedicated environment
Regulatory frameworks
Hallucinations
QuestFeed builds AI-powered decision support systems that turn complex, interconnected data — regulations, threat intelligence, technical specifications — into scored, navigable intelligence networks. Our platforms use graph-based reasoning with Bayesian probability layers to model how risk cascades through real systems.
Our AI is built on production machine learning models, Bayesian networks, and graph algorithms — not language models. LLMs handle document ingestion and communication only. The graphs reason, the models score. That means no hallucinations at the decision layer, deterministic outputs, and a fraction of the running cost of LLM-dependent systems.
Three platforms are live with self-serve sign-up: AuditROI (auditroi.com) for cybersecurity intelligence and attack chain analysis, AuditDSS (auditdss.com) for regulatory compliance intelligence across 320 frameworks, and ZoneDSS (zonedss.com) for planning intelligence with persona-specific reports for any Australian address.
Cascade risk models how a single failure propagates through an interconnected system. In cybersecurity, vulnerabilities chain into multi-step attacks. In regulation, a single compliance failure triggers enforcement across dozens of dependent rules. Our engine models these cascades using Directed Acyclic Graphs with Bayesian reasoning on every edge.
QuestFeed Pty Ltd is headquartered in Queensland, Australia. We serve clients globally, with particular focus on Australia, the United States, the United Kingdom, and Canada.
Interested in our platforms, enterprise licensing, custom integrations, or partnership opportunities? We'd like to hear from you.
Company
QuestFeed Pty Ltd
ABN
58 632 013 855
Founder & CEO
Dr. Ali Ershadi