AI Governance
StartupToGo uses AI to give students feedback and drafting help across the 5-module entrepreneurship journey. Here's exactly which models we use, what data reaches them, our guardrails, and where a human stays in control.
Models we use
Every AI feature calls the OpenAI API from our server — never the browser, and the API key is never exposed to the client.
- Idea evaluation & scoringOpenAI · gpt-4o
- Business Model Canvas review & prefillOpenAI · gpt-4o
- Pitch deck generationOpenAI · gpt-4o
- Customer-survey insightsOpenAI · gpt-4o
- In-app AI helperOpenAI · gpt-4o-mini
What student data reaches the AI
Prompts contain only the business-idea content a student authors — the problem, solution, target customer, business model, canvas, pitch, and survey answers.
Never sent to any AI model
Student names, student emails, teacher names, school or cohort names, certificates, or any login identifier. Survey respondent names are stripped on our server before the analysis call — the AI sees the answers, not who gave them.
Guardrails
- PII minimisation — identity fields are never assembled into a prompt; respondent names are removed at the boundary.
- Keys stay server-side — the AI provider key is never in the browser.
- Provider data use — per OpenAI's API policy, data sent via the API is not used to train their models and is retained only briefly for abuse monitoring, then deleted.
- No silent automation — AI output is advisory. It never changes a grade, issues a certificate, or takes an irreversible action on its own.
A human stays in control
- AI scores and suggestions are feedback — the student decides what to act on.
- Teachers review every student artifact and leave comments.
- A certificate is issued only when a teacher approves a graduation request. There is no auto-graduation.
Questions or concerns
Email Belal.riyad@startuptogo.app.