Human factors consulting
Find the behaviours, incentives and decision points that make otherwise robust security controls fail.
CREST Certified in Cybersecurity Training
Security,
made intelligible.
SanRa helps organisations secure the human and artificial intelligence systems they depend on.
Talk to us- 01
- Human factors
consulting - 02
- ISO
compliance - 03
- AI security
scanning
Consultancy 01 / 04
Build confidence
into complex systems.
Technology succeeds only when people can use, challenge and govern it with confidence. Our work joins behavioural insight with technical cybersecurity.
ISO compliance
Practical readiness for ISO/IEC 27001, ISO/IEC 42001 and related management-system programmes.
AI security scanning
Probe AI systems and tooling for prompt injection, leakage, malicious models and rogue agent behaviour.
The SanRa lens 02 / 04
Every system has
a human perimeter.
Security is not a feature you bolt on. It is a relationship between a system, its people and the decisions they make.
01Understand the system
02Expose the weak signals
03Design defences that last
AI security self-assessment 03 / 04
What is your
AI risk signal?
Three questions. A useful starting point. No data is retained unless you choose to contact us.
Your signal
Open the FIDO framework Open the ISO 27001 framework Open the ISO 42001 framework Discuss your result with a specialistThis indicative assessment is educational only. It is not an audit, certification or compliance determination.
Insight 04 / 04
FIDO
Passkeys that
survive phishing.
FIDO2, WebAuthn and passkeys — policy, recovery, IdP rollout and assurance evidence in one interactive framework.
Open the frameworkISO/IEC 27001
An ISMS people
can actually run.
Interactive clause framework for information security management — scope, risk, SoA and operational evidence.
Open the frameworkISO/IEC 42001
AI management,
made ready.
Decisions, evidence and operating disciplines behind an AI management system — plus an interactive clause framework.
Open the framework Download readiness guide ↓WP-01
The Neuron Doctrine
of Security.
SanRa thought leadership on human factors and security in the age of AI-augmented systems.
Download whitepaperProduct
Bastion — keep AI from being jailbroken.
Closed-source safety library: jailbreak heuristics, ethics policy packs, and output gates. Pair with open-source Open Bastion.
Product
Botglass — bots meet Cloudflare.
Score bot risk from Cloudflare bot scores and request signals. Studio plus an open Worker starter.
Product
Verisight — deepfake signal scoring.
Heuristic synthetic-media risk from EXIF gaps, artefacts, and claim conflicts. Honest limits, local studio.
Product
Converge — SASE-style policy plane.
ZTNA, SWG and CASB decisions in one studio — SanRa’s take on the converged edge category (inspired by platforms like Cato; not affiliated).
Product
Embodied Cognition — reason in body-space.
Brain-inspired spatial reasoning for robots: Somatic Place Graphs and hippocampal affordance replay, with a local Embodiment Bus so tools join as senses and effectors.
Product
Taste — get into the mind of the market.
Model three minds: North Korean hacker, Russian hacker, and the usual consumer. Rank messaging that hits — and the breaches they’d sniff for, with or without AI.
Product
White Owl — see the city, find the cracks.
Free WhiteHat desktop app for AI-era codebases. Visualise architecture as a 3D code metropolis and surface vulnerability signals — locally, on your machine. Not open source.
Open source
Drillwright — train specialised security SLMs.
Collect discipline-scoped cybersecurity data and run local inference against Foundation-Sec, Antares, and other OpenAI-compatible endpoints. MIT licensed.
Open source
Open Bastion — jailbreak heuristics, MIT.
Open subset of SanRa’s AI-safety prompt gating. For the full ethics packs and studio, use closed-source Bastion.
Open source
Anti-Body — your immune system for the AI age.
Free, offline AI-security scanner: prompt injection, secrets leakage, malicious models, phishing lures, and rogue agent actions. MIT licensed. Your data never leaves your machine.
Open source
Botglass Worker — Cloudflare bot starter.
MIT Pages/Workers middleware patterns for basic bot friction. Pair with proprietary Botglass for full scoring.
Training
Two free courses for the AI age.
SAN-101 builds judgement with AI in the loop. SAN-102 closes the data risks of using AI. Free online for now — not CPD or (ISC)² accredited.
Credentials & collaboration
Credentials
CREST Certified in
Cybersecurity Training
Small print
EC-Council & Proofpoint certified
Guest lecture at SANS
Partners
Erasys Ltd.
TAAM Open Source
Anti-Body · Drillwright · Open Bastion
Read our privacy notice.