Humanix
A San Francisco startup defending the 'human layer' with conversational AI that detects social-engineering and voice-impersonation attacks in real time.
Visit Website ↗ + Add to Compare Claim This CompanyOverview
Humanix builds what it calls a Human Threat Detection and Response (HTDR) platform: conversational AI trained on human psychology and manipulation tactics that monitors voice, chat, email, video, and help-desk interactions for signs of social engineering, impersonation, and deception. Rather than instrumenting endpoints or networks, it analyzes the conversation itself, flagging urgency tactics, authority claims, and policy-evasion attempts as they happen, then guides employees or help-desk staff with real-time intervention steps.
Founded in San Francisco in 2023 by CEO Keith Stewart (previously of Cisco, Brocade, Riverbed, and vArmour), the company raised roughly $18 million across combined seed and Series A rounds led by BoldStart Ventures and Acrew Capital, announced in November 2025. It was named a Top 10 finalist in the RSA Conference 2026 Innovation Sandbox contest, a juried recognition that lends independent credibility to an otherwise young, thinly-evidenced product category.
Innovation Matrix Assessment
Introduced a genuinely new product framing (HTDR) and shipped a working real-time conversational-AI detection pipeline covering voice, chat, email and video within roughly two years of founding.
Targets help-desk social engineering and voice/video impersonation, a documented and growing attack path that legacy endpoint and awareness-training tools don't directly address.
Raised $18M from credible seed/Series A investors and was independently selected as an RSAC 2026 Innovation Sandbox Top 10 finalist, though customer counts remain undisclosed.
Positions itself as a new category (human threat detection and response) rather than an incremental add-on to DLP or security-awareness training.
No independently verified customer case studies or third-party test results were found; the RSAC finalist slot is a credibility signal, not proof of real-world efficacy.
AI-generated voice/video impersonation and social engineering are growing, durable threats that will matter more, not less, over the next several years.
Why CISOs Should Care
Gives CISOs a way to catch social-engineering and voice-deepfake attacks on help desks and employees in the moment, rather than only after the fact.
What Makes It Different
Analyzes the conversation and behavior itself using psychology-trained conversational AI, instead of relying on endpoint, network, or static awareness-training signals.
The Matrix Verdict
67/100 — INCREMENTAL INNOVATOR
Humanix is an Incremental Innovator: a well-funded, independently recognized early mover in a genuinely new category, held back from a higher tier only by the absence of disclosed real-world efficacy evidence.
Editorial Note: Claims vs. Verified Findings
Funding and RSAC finalist status are independently reported; specific detection-accuracy or deployment-scale claims come only from the company and have not been independently verified.
Sources
Alternatives to Humanix
Abnormal AI
AI-native behavioral security platform that analyzes sender identity and communication patterns, rather than message content alone, to stop…
Palo Alto Networks Cortex XSIAM
Palo Alto Networks' AI-driven 'autonomous SOC' platform that unifies SIEM, EDR, SOAR, and attack-surface data into a single…
7AI
7AI, founded by Cybereason's former CEO and CTO, deploys autonomous AI agents that triage, investigate, and correlate security…
Sevii
CISO ReviewedAgentic AI platform positioning itself around autonomous threat elimination rather than alerting, using AI agents to investigate and…
Fenix24
Chattanooga-based ransomware recovery specialist that has restored operations after 500+ real-world incidents, including 30 Fortune 500 companies.
Artemis Security
A New York AI-native SOC platform correlating identity, network and cloud signals into attack narratives, already counting Wix…