Before you submit anything
Use public, synthetic, anonymised or non-sensitive material. Do not enter passwords, API keys, customer records, health information, financial details, identity documents, private contracts, legally privileged material, security findings or information your employer or client restricts.
How the experience works
Your instructions and relevant context are sent to cloud and AI providers to produce the requested result. Different tools may use text models, image models, public website extraction or document-processing services. NeoSpark applies structured prompts, guardrails, output checks, usage limits and human-readable warnings, but these controls cannot eliminate every error or risk.
AI can make mistakes
- Facts, calculations, citations, summaries and labels may be incorrect.
- Generated designs may miss accessibility, security, privacy or regulatory requirements.
- Generated code and architecture must be tested and security-reviewed before production.
- Generated content can resemble third-party material and may not be unique.
Your review checklist
- Verify important facts and source claims.
- Remove synthetic data and replace it only through an approved process.
- Review privacy, security, accessibility and legal obligations.
- Test workflows, calculations, code and integrations.
- Have an accountable person approve the result before use.
What NeoSpark records
We record privacy-preserving operational statistics such as the tool used, processing time, success or failure, model and token totals. Those aggregate logs do not contain the full prompt. A full prompt may be retained with a seven-day artifact pack only when you expressly select the separate improvement-review option. See the Privacy Notice for details.
Questions or concerns
Stop using the output if it appears unsafe, discriminatory, misleading or technically unsound. Contact privacy@neospark.com.au for data concerns or hassan@neospark.com.au for product and AI-governance questions.