Most Cover Letter Tools Write for You. Almost None Think for You.

The cover letter tool market has exploded. AI generators, resume-to-letter converters, template builders with smart suggestions. Every major career platform now offers some version of "paste your resume, paste the job description, click generate."

After researching how these tools actually work — not their marketing, but their methodology — we found something surprising. Beneath the AI branding, almost all of them are doing the same thing.

The Matching-and-Polishing Model

Here is what most cover letter tools actually do:

First, they scan your resume for keywords and skills. Second, they scan the job description for keywords and requirements. Third, they match the overlapping terms and generate text that connects them. Fourth, they polish the output: clean grammar, professional tone, proper formatting.

Some do this better than others. Some use GPT-4. Some use proprietary models. Some let you choose between "formal" and "creative" tones. But the underlying logic is the same: keyword extraction, keyword matching, text generation, polish.

This is, essentially, the standard advice from Part 1 of this series — automated. The three-paragraph model, the keyword-matching method, the template-plus-customisation approach — turned into software.

And for the same reasons the standard advice works, these tools work. They produce competent cover letters faster than most people can write them manually. They ensure keyword alignment. They eliminate typos and formatting errors. They save time.

But they also inherit the same limitations.

Writing vs. Thinking

The fundamental question these tools answer is: "How should I say this?"

The question most applicants actually need answered is: "What should I say?"

There is an enormous difference. A tool that helps you write more clearly is valuable. But if you are writing the wrong thing clearly, clarity does not help.

Consider what happens when you paste a job description into a typical generator. The tool extracts keywords — "project management," "cross-functional collaboration," "data-driven decision making." It finds these terms (or synonyms) in your resume. It generates sentences connecting them: "In my role at X, I demonstrated strong project management skills through cross-functional collaboration..."

The output is grammatically perfect. It is keyword-aligned. It will pass an ATS. But it has not done any of the hard work:

It has not figured out what the recruiter is actually worried about — the unstated problem that created this job opening.

It has not identified which of your dozens of experiences is the one that speaks to that specific worry.

It has not structured the narrative to answer the recruiter's unspoken questions in the order they will ask them.

It has not decided how to handle the fact that you are switching industries, or that your last role was two levels below this one, or that your strongest qualification is in a domain the job posting never mentions.

The tool wrote. It did not think.

The Four Competencies, Revisited

In an earlier post, we described the four things a cover letter actually requires: understanding what the recruiter needs (which we called HR intelligence), understanding which parts of your experience are relevant (self-reflection), structuring the argument persuasively (narrative craft), and executing the whole thing technically (orchestration).

Most tools handle orchestration well. Some handle narrative craft at a basic level — they can structure three paragraphs and choose professional language. Almost none handle HR intelligence or self-reflection.

This is not a criticism of the engineering. It is a reflection of what these tools were designed to do. They were built to automate the writing process. The thinking process — the strategic layer that determines what gets written — is left entirely to the user.

Which means the user still needs to be the strategist. They still need to know which achievement to highlight. They still need to understand what the recruiter cares about. They still need to decide how to position their gaps and weaknesses. The tool just types faster.

When Writing-Quality Tools Are Enough

To be fair, there are situations where a writing-quality tool is exactly what you need.

If you already know what to say and just need help saying it well, a grammar-focused tool is perfect. If you are a strong communicator applying to a role you deeply understand, a template builder saves time without sacrificing quality. If you need to produce one cover letter for a job you are highly qualified for, the standard approach works fine.

The tools are not bad. They are just solving a specific problem — and that problem is not the one most applicants actually have.

The Harder Problem

The harder problem is the one from Part 2: most people do not know what the recruiter needs to hear. They cannot objectively evaluate which of their own experiences is most relevant. They default to their proudest achievement instead of the most strategically useful one. They hide gaps instead of positioning them. They apply the same approach across twenty applications and wonder why the results decline.

A tool that solved this problem would not just match keywords. It would analyse the job description the way a recruiter reads it — not just extracting requirements, but understanding what they signal about the company's challenges. It would not just scan your resume — it would understand what your experience actually means in context, including the connections you cannot see yourself. It would not apply one narrative structure to every letter — it would choose the right structure based on the specific relationship between your background and this specific role.

It would not just write for you. It would think for you.

We Know This Because We Tried

We spent two years building exactly this kind of system. The thinking part turned out to be much harder than the writing part. We got it wrong many times before we got it right. We learned that the gap between keyword matching and genuine strategic intelligence is not a feature gap — it is a philosophical gap.

If you are curious about what that looks like in practice, the first letter is free.

This is Part 3 of a three-part series on what the internet teaches about cover letters. Part 1: What They All Agree On. Part 2: Where the Standard Advice Stops Working