How our AI thinks, in plain English.
Every AI feature in the Career Operating System is designed so you can see what it did, why it did it, and what it did not do. This page explains the models, prompts, and safety layers behind each capability.
What the AI does
- Resume analysis — extracts skills, structure, keywords, and scores your resume against a target role.
- ATS scoring — deterministic, rule-based scoring against a target job description. Every point is explainable.
- Career roadmap — composes a 90-day / 6-month / 1-year plan from your profile and goals.
- Interview coaching — generates questions, evaluates answers across 9 rubrics, and returns coaching feedback.
- Application intelligence — helps track and prioritise, and suggests follow-up timing.
- Offer comparisons — benchmarks against public market data and negotiation heuristics.
What the AI does not do
- It does not apply for jobs on your behalf without explicit action.
- It does not impersonate you in interviews or written communications.
- It does not share your data with recruiters unless you explicitly opt in.
- It does not make hiring decisions on behalf of employers.
Models we use
The platform is provider-agnostic. Depending on task and cost, we route to leading commercial models (OpenAI, Anthropic, Google) via a unified adapter. You can see the model used for any AI response inside the platform under the response’s footer.
Human oversight
Every AI action is reviewable by you. Every generated document is editable. Every recommendation includes a “why” card so you can accept, edit, or dismiss it.
Reporting a problem
Found an output that feels unfair, wrong, or unsafe? Report it via Support or the in-app feedback widget. We review every report within one business day during Beta.