Assisted tasks
Classification, extraction, summaries, eligibility analysis, project matching, drafting and contradiction checks may use AI.
Responsible AI
AI supports institutional judgment; it does not replace it.
Growth OS uses AI to organize, compare and draft from controlled context while keeping human review and source verification visible.
Human approval
No application, email, decision or publication is sent automatically from an AI draft. An authorized user must review and approve it.
Sources and confidence
Growth OS distinguishes official sources, imported data, unverified fields, inference and approved institutional claims.
Provider
When enabled, selected prompts and context are sent through the OpenAI API. OpenAI states that API business data is not used for model training by default unless the customer opts in.
Known limitations
AI can misunderstand eligibility, dates, legal status, amounts or context. Scores and recommendations are not guarantees.
Auditability
Where available, Growth OS records the model, operation tier, token usage, source context, date, user and approval history.
Guardrails
Rules built into the product
- Do not invent partners, results, budgets or legal status.
- Do not convert thematic affinity into eligibility.
- Keep historical facts separate from projections.
- Use approved claims and retain supporting evidence.
- Escalate legal, tax, safeguarding and high-impact decisions to qualified professionals.
Need to review a specific AI workflow?
Request a demonstration focused on sources, eligibility, scoring, drafting and human approval.