Legal

AI Use Disclosure

Last updated: 11 May 2026

GradProcess AI uses artificial intelligence to provide feedback, analysis, and coaching. This page explains exactly how AI is used, what data is processed, and its limitations.

Model Used

Anthropic Claude (claude-3-5-sonnet / claude-3-opus)

Data Sent

Content only — never your name or email

Limitations

Guidance only, not a guarantee of outcomes

1. AI Provider

GradProcess AI uses Anthropic's Claude language models, accessed via the Anthropic API. Anthropic is a US AI safety company. Their data processing terms apply to content sent to their API. Anthropic does not use API inputs to train their models by default.

Anthropic Privacy Policy →

2. Where AI Is Used

CV Tailoring & ATS Scoring

Data sent to AI

Your CV text and the job description you provide

AI output

ATS score, keyword gaps, tailored bullet suggestions

STAR Builder

Data sent to AI

The situation/task/action/result text you write

AI output

Structured STAR feedback, suggested improvements

Video Interview Coaching

Data sent to AI

Your spoken transcript (audio-to-text) and interview question

AI output

Content score, delivery feedback, STAR structure rating

Application Tracker — AI Breakdown

Data sent to AI

Company name, role title, sector, and job description text

AI output

Role summary, interview questions, prep roadmap, readiness score

Application Tracker — Role Extraction

Data sent to AI

Job description text you paste in

AI output

Structured role fields (company, sector, skills, deadline)

Ava — AI Coach

Data sent to AI

Your typed message and short conversation history

AI output

Coaching guidance and answers to preparation questions

Industry Hub — AI Insights

Data sent to AI

Your selected sector

AI output

Commercial themes, watchlist, talking points for interviews

3. What We Never Send to the AI

Your name, email address, or account identifiers
Your payment information
Other users' data
Passwords or authentication credentials
Raw video or audio files (only transcribed text is sent)

4. Prompt Injection Protection

User-supplied content (CVs, job descriptions, transcripts) is wrapped in explicit boundaries before being sent to the AI. This instructs the model to treat the content as data only and ignore any instructions that might appear within uploaded documents.

We also apply content length limits and output validation to reduce the risk of unexpected AI behaviour.

5. AI Limitations — Important

Please read carefully

  • AI scores and feedback are guidance only — not a prediction of actual interview or application outcomes
  • Body language analysis is experimental and may be inaccurate
  • ATS scores are estimates based on keyword matching — actual ATS systems vary widely
  • AI may hallucinate company details or role information — always verify independently
  • Interview questions suggested by AI are examples only, not guaranteed to appear
  • Do not rely solely on AI feedback — seek guidance from careers advisers and mentors

6. Human Review

GradProcess AI does not employ human reviewers who read your CV, recordings, or practice sessions. All feedback is generated automatically by AI models. You should treat all AI output as one data point, not as definitive professional advice.

7. Data Retention by Anthropic

By default, Anthropic does not retain API request data beyond 30 days for trust and safety monitoring. They do not use API data to train their models without explicit consent from API customers. GradProcess AI has not granted Anthropic permission to use your data for model training.

8. Your Controls

  • You can disable AI feedback for practice sessions in the recording consent modal
  • You can delete your recordings and transcripts at any time
  • You can export all your data including AI-generated content via Settings → Privacy

Questions about our AI use? Email us at ai@gradprocess.ai. See also our Privacy Policy and Terms of Service.