You paste a job posting. You upload your CV. You click a button. Three minutes later, you have a cover letter that sounds like you wrote it on your best day, tailored to a job you found twenty minutes ago.

It feels simple. But what happens in those three minutes is anything but. And understanding it helps explain why the result is so different from what you'd get by asking ChatGPT to "write me a cover letter."

Step 1: Reading the Job Posting Like a Recruiter

What you'd have to do manually: Read the posting three or four times. Highlight the key requirements. Google the company. Try to figure out what they really care about versus what's just boilerplate. Research their products, competitors, recent news. Piece together what problem they're actually hiring someone to solve.

Most people skip most of this. They skim the requirements, check if they roughly match, and start writing.

What StoryLenses does instead: Our AI — powered by Anthropic's Claude Sonnet 4, one of the most capable reasoning models available — reads the entire posting and extracts over 15 structured data points. Not just the obvious ones like job title and required skills, but the things a seasoned recruiter would notice: What business problem is driving this hire? What are the company's current challenges? What does their language reveal about their culture?

It reads between the lines. A posting that says "fast-paced environment" and "wearing many hats" tells us something different from one that says "established processes" and "cross-functional alignment." These signals shape the story we'll tell about you.

Step 2: Understanding Your CV Better Than You Do

What you'd have to do manually: Sit with your CV and try to recall the details of every role. That project three jobs ago where you led a team of twelve — what was the business impact? Those skills you use every day but never think to mention — how do you even name them? And if your CV is in Portuguese but the job is in German, you're doing all of this translation work on top of everything else.

What StoryLenses does instead: You upload a PDF — or even a screenshot of your LinkedIn profile — and the AI extracts everything. Complete work history with no truncation. Skills sorted into categories: soft skills, hard skills, technical skills, domain expertise. Achievements pulled out with their metrics intact. Languages, certifications, education — all structured and ready to match.

The key insight: it often surfaces connections you didn't see yourself. That marketing role where you managed vendor relationships across three countries? That's international stakeholder management. The AI names it, even if you didn't.

Step 3: The Matching That Nobody Does by Hand

What you'd have to do manually: Open the job posting in one window, your CV in another, and try to draw lines between them. This is where most people fall apart. You either undersell yourself — missing valid connections between your experience and their requirements — or you oversell, claiming everything without evidence.

And here's the part almost nobody does: figuring out your gaps and proactively addressing them. What does the job require that you don't have? How do you acknowledge that without torpedoing your application?

What StoryLenses does instead: This is where the real intelligence lives. The AI performs semantic skill matching — it understands that "Projektmanagement" and "Project Management" are the same thing, that "Agile Methodologies" overlaps with "Scrum Master experience," that leading a team of 15 in São Paulo is relevant experience for a "people management" requirement in Berlin.

It calculates a match percentage, identifies your gaps, and — critically — suggests mitigation strategies for each gap. If you're missing a required certification, it might highlight equivalent experience that demonstrates the same competency.

On top of that, it analyzes your professional archetype. Are you a Bridge-Builder who connects teams? A Fixer who thrives in turnaround situations? A Strategist-Executor who both plans and delivers? Your archetype shapes how the story is told — leading with your strongest angle rather than a generic chronological summary.

Step 4: Writing in a Voice That's Yours, Not a Robot's

What you'd have to do manually: Stare at a blank page. Write a first sentence. Delete it. Write another one. Google "cover letter examples." Copy a template. Realize it sounds like every other template. Try to make it personal. Give up and send it anyway.

Or you paste your CV into ChatGPT and write "make me a cover letter." You get something grammatically correct that could have been written by literally anyone. No specificity. No personality. No story.

What StoryLenses does instead: This is where all the previous steps converge. The AI doesn't just "write a cover letter" — it constructs a narrative using 11 proven storytelling structures and 7 professional tones, each calibrated for different industries and roles.

The storytelling structures are adapted from narrative archetypes that have worked across cultures for millennia. The Golden Fleece for someone who deliberately pursued growth through certifications and stretch assignments. The Problem-Solver for someone whose strength is diving into urgent challenges. The Fool Triumphant for someone with a non-traditional background who brings a fresh perspective that traditional candidates can't.

The tones range from Analytical — data-driven, metrics-heavy, cause-and-effect reasoning perfect for finance or engineering — to Storytelling — vivid, emotionally resonant, perfect for creative or marketing roles. A cover letter for McKinsey should not sound like a cover letter for a design studio, and now it doesn't have to.

And because the AI has your matched skills, your gaps with mitigations, your archetype, and the company's specific challenges — it writes something that actually sounds like you making your case for this specific job.

Step 5: The Language Problem That Translation Can't Solve

What you'd have to do manually: If you're applying in a language that isn't your strongest, you have two bad options. Write in your native language and translate — but translations read like translations, with awkward phrasing, wrong conventions, and a tone that doesn't land. Or write directly in the target language and spend three times as long, second-guessing every word choice.

What StoryLenses does instead: The AI generates natively in your target language. This is a fundamental architectural choice, not a nice-to-have. When you select German, the entire generation process — the thinking, the phrasing, the cultural conventions — happens in German from the start. It's not an English letter run through a translator.

This matters more than you might think. A German Bewerbungsanschreiben follows different conventions from an American cover letter. The structure is different. The level of formality is different. The way you express confidence is different. When the AI thinks in German, it writes like a German professional, not like a translated American one.

The same applies to Portuguese and English. Each language carries its own professional culture, and the AI respects that.

Step 6: Making It Better, Not Starting Over

What you'd have to do manually: Read your draft. Decide paragraph two is weak. Try to rewrite it. Accidentally break the flow with paragraph three. Fix that. Now the opening doesn't work with the new middle. Rewrite the opening. You've spent another hour and you're not sure it's better.

What StoryLenses does instead: You can lock the paragraphs you love, comment on the ones you want changed, and regenerate. The AI refines surgically — keeping what works, improving what doesn't. You can do this up to five times per story, building toward a result you're genuinely proud of.

Each version is saved, so you can always go back. No work is lost. No good paragraph is accidentally destroyed in pursuit of a better one.

The Models Behind It All

We use Anthropic's latest Claude models, chosen specifically for different parts of the pipeline:

Claude Sonnet 4 handles the heavy lifting — job analysis, profile matching, and story generation. It's one of the most capable AI models available for complex reasoning and nuanced writing. When you need an AI that can understand that your experience scaling a startup in Latin America is directly relevant to a growth role at a European SaaS company — that takes real reasoning, not just pattern matching.
Claude Haiku 4.5 handles the fast, lightweight tasks — real-time explanations of narrative options as you hover over them, quick scoring calculations, and contextual tips. It responds in milliseconds, keeping the experience snappy.

Every AI call is instrumented with observability tooling, so we can monitor quality, catch issues, and continuously improve the prompts that drive the entire experience.

Why This Matters

The technology is sophisticated, but the point was never the technology. The point is this: a cover letter shouldn't take longer to write than the interview preparation itself.

You have the skills. You have the experience. You have the drive that made you apply in the first place. What you shouldn't need is three hours of agonizing over how to say "I'm good at this job" in a way that doesn't sound like everyone else.

We built StoryLenses so your skills do the talking. We just help you find the right words.

Blog

After the Interview Is Before the Interview — How to Use Interview Insights to Refine Your Narrative