Independent Pre-Launch Review
A sober, evidence-based review of AI-built, AI-assisted, vibe-coded, or rapidly prototyped digital materials before wider use or launch.
The purpose is to help businesses and organisations understand whether an app, website, chatbot, form, report, internal tool, public-facing service, AI assistant, or digital workflow is clear, trustworthy, privacy-conscious, source-grounded, resilient, and ready for wider use.
We review AI-built or AI-assisted apps, websites, forms, reports, internal tools, public-facing materials, and digital workflows for practical launch readiness.
Checks may cover UX clarity, misleading claims, source accuracy, legal and compliance caveats, privacy and data minimisation, accessibility basics, security red flags, governance gaps, and GDPR and EU AI Act compliance readiness considerations.
For AI chatbots and assistants, we can include scoped LLM resilience testing: low-volume, manual testing of behaviour under realistic and adversarial user prompts.
This includes prompt-injection resistance, hallucination and source-grounding checks, unsafe advice boundaries, privacy and sensitive-data handling, overconfident legal or compliance claims, refusal and escalation behaviour, consistency across repeated or reframed questions, and handling of long, malformed, hostile, or ambiguous prompts.
We agree the review scope, create a test matrix, carry out manual review, capture screenshots and notes, classify findings by severity, and provide practical recommendations.
The output is a concise report with prioritised findings and concrete next steps in formats suitable for human review and LLM-assisted follow-up. A human consultant can support interpretation and implementation as needed.
This is practical independent review, not certification, legal advice, a guarantee of compliance, or penetration testing.
LLM resilience testing is not high-volume load testing, endpoint probing, model extraction, provider probing, credential or admin testing, or invasive security testing unless separately agreed and authorised.
We do not ask clients to share unnecessary personal, confidential, or sensitive data.
A focused check for early-stage materials or a limited workflow.
Best when you need a quick independent read on clarity, obvious risks, user-facing claims, and next-step priorities.
A broader review before wider internal use or public launch.
Best for apps, websites, forms, reports, internal tools, public-facing materials, or AI-assisted services that need a structured readiness check.
Manual behavioural testing for chatbots and AI assistants.
Best when a chatbot or assistant needs evidence on source grounding, refusal behaviour, privacy handling, prompt-injection resistance, and boundary-setting.
A follow-up check after changes have been made.
Best when you want to confirm whether priority recommendations have been addressed before launch or wider rollout.
Send a short description of the material, who will use it, what stage it is at, and any known privacy, compliance, or trust concerns. We will suggest a proportionate review scope.
Book a review before launch