What AI Resume Tailoring Actually Changes (A Real Before/After)

The honest worry about AI resume tools is that they'll invent experience you don't have — a fabricated metric, a skill you've never touched, a title you never held. It's a fair worry, and the only real way to address it is to show the actual output rather than just claim it doesn't happen.

So here's a real one: a test resume run through Wonzo's AI Resume Tailoring against a live Software Engineer Internship posting at a real fintech company (Clerkie), pulled straight from their actual Greenhouse listing.

What the job posting actually asked for

The real posting described a full-stack role built on Next.js, Tailwind CSS, Node.js, AWS, and MongoDB, working on a B2B CRM product for a fintech company, with a strong interest in candidates who could speak to financial-wellness or fintech context specifically.

What the tailored output actually said

The generated summary read: "Software engineer with hands-on experience across the full stack, skilled in JavaScript/TypeScript, Node.js, and REST API development, paired with cloud and DevOps fundamentals (AWS, Docker, CI/CD). Comfortable working with relational and in-memory data stores (PostgreSQL, Redis) and microservices architectures, with a strong interest in building product-focused, full-stack features similar to Clerkie Fiber's B2B CRM platform."

Every technology named in that summary — Node.js, AWS, Docker, PostgreSQL, Redis — was already on the source resume. Nothing was invented. What changed was framing and emphasis: the summary explicitly names the company's own product ("Clerkie Fiber's B2B CRM platform") and leads with full-stack/cloud experience because that's what this specific posting asked for, not because the underlying resume changed.

The output also included a suggestedKeywords list — Next.js, Tailwind CSS, MongoDB, fintech, B2B CRM — pulled directly from the job description. These are deliberately kept separate from the tailored summary rather than woven in as if the candidate already had them, because the source resume didn't list them. That's the actual mechanism: reorder and reword what's true, and hand back what's missing as a suggestion instead of quietly fabricating it into the resume itself.

Why it works this way on purpose

The system prompt behind Wonzo's tailoring feature is explicit about this boundary: reorder, rephrase, and emphasize — never invent experience, skills, projects, or achievements the resume doesn't already show. If something in the job description has no basis anywhere in the candidate's actual resume, it goes into a suggestions list for the candidate to decide on, not into the resume itself. That's a harder constraint to build than just generating a resume that sounds better, and it's the difference between a tool you can actually trust with something as consequential as a job application and one that quietly puts words in your mouth.