Explainable Decision Cycle
Six product features support Human-AI Teaming in the decision support cycle
Five AI advisory products across Industry, Finance, and Defense. Every recommendation is explainable, every causal path is visible. Start with SOP Builder — find the gaps in your procedures before an incident does.
Diagnose and fix low-flow problems in artificial lift wells. Expert knowledge encoded once — scales globally without the expert present.
Identify complex VAT fraud schemes across 2.5M+ daily transactions with auditor-grade explainability. 7% annual revenue uplift for three consecutive years.
Detect and manage grid situations in real time. 43,000 situational assessments per day with full explainability for operators and regulators.
Monitor 60,000 traders every 4 seconds. Detect suspect positions with investigator-ready causal explanations — not just a risk score.
Assure quality of industrial ultra-purified water. 3,000 control points assessed every 30 seconds with real-time cognitive oversight.
Convert existing documents into validated SOP Maps. Simulate scenarios before go-live. Find gaps before incidents do — in weeks, not months.
CRex™ is HMX's patented Cognitive Reasoning Engine. Causal, deterministic, goal-directed — not a language model, not a black box. Every recommendation traces to verifiable data through an explicit causal chain operators can read, challenge, and defend in any mission-critical environment.
Every CRex agent works toward explicitly defined enterprise goals — not pattern matching. Outcomes are reliably deterministic and goal-aligned.
Situation-centric reasoning understands context and responds dynamically to changing environments.
Every recommendation includes natural language reasoning. No black boxes — ever. Every outcome is auditable and regulator-ready.
Causal Bayesian inference handles real-world uncertainty and incomplete data — mirroring how experts reason under ambiguous conditions.
Chain-of-thought reasoning orchestrates multiple AI models, third-party agents, and domain logic into a single unified cognitive agent.
Supercharges enterprise IP by unifying expert knowledge with deep ML. Expert knowledge encoded once becomes infinitely scalable.
Encode expert knowledge — goals, situations, and assessor formulas — into structured, auditable reasoning models that grow richer over time.
Upload documents, SOPs, and domain knowledge. CAB instantly recommends a SOP Map. Compress deployment from months to days.
Validate SOP Maps and run scenario simulations directly in Excel. No IT required.
Production deployment at scale. Multi-agent architecture for high-velocity transactions and real-time industrial control. Cloud and on-premise.
Ceará had 60,000+ POS devices generating 2.5M+ daily transactions. HMX delivered a Tax Intelligence System fusing expert VAT logic, risk rules, ML models, and Bayesian causal inference — trusted by auditors and ministry executives.
HMX encoded Rockwell's world-renowned production specialist expertise into a CRex SOP Map — diagnosing low-flow conditions using unified causal reasoning across rules, domain models, and Bayesian ML.
HMX delivered the Foreign Exchange Intelligence System — processing real-time trading data every 4 seconds across 60,000 traders with continuously learning Bayesian assessors and investigator-ready explanations.
HMX exists to make AI explainable — fully, verifiably, and without exception. We build cognitive products that empower enterprises to achieve operational excellence with human-like reasoning that is reliable, transparent, and accountable. Every recommendation our systems make comes with a complete causal explanation. No scores without rationale. No alerts without context. No black boxes.
We believe the black box model of AI is not just a technical problem — it is a trust problem. In mission-critical environments, unexplained AI is unusable AI. HMX was founded on the conviction that the most powerful AI is the AI people can read, challenge, and trust.
Every HMX product surfaces the full causal chain behind every recommendation — in plain language operators and regulators can act on and defend.
We build against the grain of opaque AI. CRex™ is deterministic, goal-directed, and auditable by design — not as an afterthought.
Our cognitive products are built to amplify human expertise — not replace it. Human + AI is more capable than either alone.
Kim Mayyasi and Mike Barnett met at one of the earliest MIT artificial intelligence spin-outs, Gensym. As CEO and EVP of Solutions respectively, they collaborated on hundreds of innovative applications for industry and government — including NASA, DoD, process control, and supply chain use cases.
Ten years ago, the two joined forces again with HMX.ai. Their focus on mission-critical applications using newer AI technologies — semantic web, machine learning, and now generative AI — has culminated in the CRex™ cognitive engine: a platform that fulfills their founding conviction that Human + AI is better than either alone.
Decades of AI leadership across industrial, financial, and government missions. Co-founded HMX after building transformative AI applications at Gensym (MIT spin-out).
LinkedIn →
Architect of the CRex™ cognitive engine. Thirty years engineering mission-critical AI systems for NASA, DoD, and global industrial enterprises.
LinkedIn →Tell us about your operation and where explainable AI could help. We’ll get back to you directly.
Find the gaps in your procedures before an incident does.