How Lihv Assist works, what it is built on, the standards it follows, and what it will and will not do.
Our commitment:Lihv Assist is designed to support safe, compliant aesthetic practice in the UK. Every aspect of how it works — from the AI model we chose to the governance rules built into its instructions — reflects that commitment.
Lihv Assist is an AI-powered governance and education tool built exclusively for verified, login-authenticated aesthetic practitioners on the Lihv Connect platform. It is designed to support practitioners with:
Important: Lihv Assist is not a diagnostic tool, prescribing service, or substitute for clinical training, registered supervision, or individual professional judgement. It does not replace your obligations as a registered practitioner.
All clinical decisions remain your sole professional responsibility. Lihv Assist supports your practice — it does not lead it.
Lihv Assist is powered by Claude Sonnet, developed by Anthropic — one of the world's leading AI safety companies. Anthropic was founded with a specific focus on building safe and interpretable AI systems.
We chose Claude over other AI models specifically because of Anthropic's approach to safety, their published responsible scaling policy, and Claude's ability to follow complex, domain-specific governance instructions consistently.
The Claude API operates under strict data handling terms. Your messages are not used by Anthropic to train their models.
Lihv Assist operates under a bespoke multi-section system prompt — a set of governing instructions written specifically for UK aesthetic practice. This prompt is private, held server-side, and never exposed to the client browser or any user.
The governing framework covers:
Lihv Assist is designed and governed in alignment with guidance from the following UK bodies. Practitioners should always verify current guidance at point of use, as regulatory standards may change.
Code of Practice and governance framework for UK aesthetic practitioners
Evidence-based clinical and practice standards for non-surgical treatments
POM governance, botulinum toxin, hyaluronidase, Yellow Card reporting
Guidance on cosmetic interventions and professional standards for doctors
Standards for nurse practitioners including V300/NMP in aesthetics
CAP Code and ASA guidance for social media and cosmetic advertising
UK GDPR and Data Protection Act 2018 compliance
Governance documentation structured to support CQC inspection readiness
All AI conversations are processed server-side. The Anthropic API key and system prompt are never exposed to the browser. Conversation content is retained for 90 days then automatically deleted. Metadata is retained for platform analytics.
Practitioners must not enter patient-identifying information into the Lihv Assist chat. The system is configured to detect and flag such information if submitted.
Full details of how your data is handled are set out in our Privacy Policy.
Lihv Assist is configured with content safety controls that detect and respond appropriately to harmful, dangerous, or out-of-scope content. This includes self-harm and crisis indicators, requests related to harming others, illicit drug queries, explicit content, and attempts to extract or override the governance framework.
Interactions that trigger safety controls are flagged for admin review. Where a practitioner appears to be in distress, Lihv Assist provides appropriate crisis resources including the Samaritans (116 123) and emergency services (999).
The Lihv Assist governance framework is reviewed periodically against the latest guidance from JCCP, CPSA, MHRA, and other relevant bodies. When regulatory changes occur — such as updates to the aesthetics licensing framework under the Health and Care Act 2022 — the system prompt is updated accordingly.
We maintain a version history of the governance framework so we can demonstrate what standards were in place at any given date.
Lihv Assist is an AI system. Despite the governance controls in place, it may occasionally produce responses that are incomplete, imprecise, or that do not reflect the most current regulatory position. You are always responsible for verifying any output against current UK standards and your own professional judgement.
We are committed to continuous improvement and welcome feedback from practitioners at contact@nodatech.co.uk.
Noda Tech Ltd. Registered in England and Wales. This AI Governance and Standards document was last updated in April 2026. It is reviewed periodically and updated to reflect changes in UK regulatory guidance.