How We Use AI
We built greatfit.ai on a simple belief: AI should amplify your strengths, never replace your voice. This page explains exactly what our AI does, what data it touches, and the lines we refuse to cross.
The short version: greatfit.aiuses AI in five specific places — not everywhere — to take the tedious parts of the job search off your plate. Every draft passes through your hands before it goes anywhere. Your career data never trains anyone's AI model and never leaves your account. The full picture is below.
We draft it. You own it.
Our Philosophy
Job searching is one of the most stressful experiences in modern life. We use AI to take the tedious, error-prone parts off your plate — analyzing job descriptions, structuring your experience, drafting tailored documents — so you can focus on what matters: preparing for the interview and landing the role.
But here is the thing we care deeply about: AI is not a magic wand. It can occasionally get things wrong, miss nuance, or phrase something in a way that does not sound like you. That is why every single AI-generated output in greatfit.aipasses through your hands before it goes anywhere. We draft it. You own it.
We have invested thousands of hours engineering our AI to be genuinely useful without being invasive or reckless with your career data. Below is exactly how each feature works, in plain English.
What AI Actually Does for You Here
Most job-search tools that "use AI" just help polish a single resume a little faster. Here's what AI actually does in greatfit.aithat you couldn't reasonably do without it.
Tailor every application — not just one
A static resume can't know which of your strengths matter to this job. AI reads the role closely, finds the parts of your career that line up, and drafts a sharper application for each opportunity. The work that used to take hours per job takes minutes.
Read the fit before you spend a day applying
Before you write a word, you see how well you match the role — your strengths, your real gaps, and a clear read on whether it's worth pursuing. No human reviewer can give you that in 30 seconds.
Practice with an interviewer that knows the role
Generic interview prep is generic. AI generates questions tuned to the specific job, the specific company, and even the specific interviewer's role in the process — recruiter vs. hiring manager vs. panel.
Get a senior set of eyes on your dossier
The Scout reads your full career profile the way a senior career coach would — flagging weak bullets, missing context, and places your story could be sharper. Brutally honest, by design.
None of this is magic. It's careful work, done faster than a human reasonably could, on every application — so you can spend your energy on the conversations that actually decide it.
Feature-by-Feature Breakdown
Every place AI touches your data, explained transparently.
Multi-Pass Dossier Ingestion
When you upload a resume, our system runs it through multiple specialized AI passes — each one designed to extract a specific type of information. One pass identifies your career timeline. Another detects evidence of leadership. Another analyzes your writing tone. Think of it as 11 expert reviewers each reading your resume for a different purpose, then combining their findings into one rich professional profile.
What we send
- The text extracted from your uploaded document (PDF, DOCX, or TXT)
- Nothing else — no filename, no metadata, no account info
What we never send
- The original file itself (we extract text server-side, then discard the file within 24 hours)
- Your email, password, or payment information
Fit Analysis
When you compare your dossier against a job description, AI evaluates how well your experience aligns with what the employer is looking for. It identifies your strengths, flags genuine gaps, and suggests a tailored resume strategy. This runs in two stages: first a deep evaluation, then a second pass that synthesizes specific proof points from your career history.
What we send
- Your professional dossier (career history, skills, achievements)
- The job description you're targeting
- Any gap responses you've provided during patching
What we never send
- Your contact information or personal identifiers
- Data from other users or other job applications
Resume & Cover Letter Generation
AI drafts a tailored resume or cover letter using your dossier, the target job description, and the results of your fit analysis. It selects the best template archetype for your situation, uses identity-aligned language (e.g., leadership verbs for executive roles, hands-on verbs for practitioner roles), and enforces strict page limits. Every claim in the document traces back to evidence from your actual career history.
What we send
- Your dossier, the job description, and your fit analysis
- Your template and formatting preferences (page length, sections to include)
What we never send
- Your email or account credentials
- Information about other jobs you're tracking
Company Intelligence
Before an interview, AI researches the company using live Google Search to pull real-time information — recent news, leadership changes, financial performance, culture signals. This ensures your interview prep is based on what's happening right now, not outdated training data. The raw research is then structured into an easy-to-digest briefing.
What we send
- The company name and role title
- A summary of the job description (limited to 4,000 characters)
What we never send
- Your personal career history (only the job context is shared for relevance)
- Any browsing data or device information
Interview Coaching
AI generates interview questions tailored to the specific role and interviewer persona, then coaches you on your practice answers. It scores your responses, highlights strengths, suggests improvements, and provides example answers — all grounded in evidence from your dossier. It's brutally honest, by design.
What we send
- Your dossier and the job description
- The interviewer persona (name, title, role in the hiring process)
- Your practice answers during coaching sessions
What we never send
- Audio or video recordings (we only process text you type)
- Data from previous interview sessions with other companies
Dossier Review (The Scout)
The 'Scout' performs a ruthless audit of your dossier — looking for redundant entries, weak bullet points, missing context, and opportunities to strengthen your narrative. It suggests specific, actionable edits. Think of it as a senior career coach reviewing your profile with fresh eyes.
What we send
- Your full professional dossier (career history, skills, achievements)
What we never send
- Any data beyond your career content — no account info, no usage patterns
Our Data Promises
Concrete commitments, not corporate platitudes.
Your data never trains public AI models
We access AI through Enterprise and Developer APIs with strict commercial terms. Google, OpenAI, and Anthropic are all legally prohibited from using data sent through these APIs to train their foundational models. Your career history stays yours.
Every AI request is authenticated
Every single API call to our AI systems requires Firebase Authentication and App Check verification. Your data cannot be accessed by other users, and requests cannot be forged from outside our app.
You always review before anything leaves
AI generates drafts. You edit, approve, and export. Nothing — no resume, no cover letter, no portfolio page — is ever sent to an employer or published without your explicit action.
Uploaded files are deleted within 24 hours
When you upload a resume for processing, we extract the text, run it through our pipeline, and then delete the original file. We don't keep copies of your documents sitting around on our servers.
How We Fight Hallucination
AI can make things up. We've built multiple layers of defense against this.
Evidence-Grounded Generation
Our resume and cover letter generators are explicitly instructed to trace every claim back to your actual career data. The AI is told: if you cannot find evidence for a claim in the dossier, do not write it. We use a formula — Action + Context + Impact — where each component must be sourced from your real experience.
We Push Back When You're Vague
When you respond to an identified gap in your profile, our AI does not blindly accept vague answers like "I have that" or "Yes, I did this." It requires specific details — tools, projects, metrics, or context — before resolving a gap. Vague affirmations are flagged as unverified and kept open.
Structured Output Enforcement
We force the AI to answer in fixed categories instead of free text, which eliminates most fuzzy, made-up output. Fit ratings must be one of exactly five labels. Gap types must be categorized as "Confirmed" or "Inferred."
Output Sanitization
All AI-generated content is sanitized before it reaches your screen. Even if an AI model were to produce something unexpected, our rendering layer strips it out before you ever see it.
What We Refuse To Do
These are the lines we will not cross, full stop.
We will never auto-apply to jobs on your behalf
AI helps you prepare. It never acts for you. You decide when and where to send every document.
We will never fabricate experience you don't have
Our generation system is grounded in your actual dossier data. It cannot invent jobs, degrees, or achievements.
We will never sell your career data
Your professional history is not a product. We make money by building great tools, not by monetizing your profile.
We will never use AI to make hiring decisions
greatfit.ai is a candidate tool. We help you present yourself. We are not in the business of judging, scoring, or filtering people for employers.
We will never share your data across users
Your dossier, your fit analyses, your interview prep — it is all scoped to your account. No user can ever see another user's data.
We will never use your data to train our own models
We don't build foundational AI models. We use the best ones available through secure enterprise APIs, and your data never leaves those sandboxed sessions.
Security At a Glance
Technical safeguards that protect every AI interaction.
Authentication
Firebase Auth on every request
App Verification
Firebase App Check prevents forged requests
Rate Limiting
Per-user limits on all AI endpoints
Input Validation
Strict size limits on all user inputs
PII Scanning
Client-side detection of SSNs, card numbers, API keys before upload
File Security
MIME type + magic number verification, active content blocking
Injection Defense
XML-delimited prompt architecture isolates user data from AI instructions
Audit Logging
Every AI call, token usage, and security event is logged