How Good Are AI Cover Letter Generators? An Editor’s Perspective (With Our Own Prompt)

To a large extent, cover letters really are an outdated tradition that only persists because no one’s ever pushed hard enough to change it.

The job market in 2026 is brutal for applicants, but many recruiters and hiring managers are swamped, too. Most simply don’t have the time to read cover letters in much depth (or at all).

However, when hiring managers do read them, here’s what they generally look for:

• Communication ability: It’s easy to write “excellent communicator” on your resume, but a cover letter gives you the chance to demonstrate your skill. If your writing is competent, that’s a signal that you’d be good at other aspects of the job that require expressing yourself.
• Professionalism: Informal conventions have evolved around how cover letters should be written. If you follow those, it indicates that you understand professional norms in general.
• Enthusiasm: One convention about cover letters is that they should include a few words about how excited you are for the job. In the real world, of course, people often apply to jobs they aren’t excited for simply because we live in a society where we need money to live, but managers still look for that sense of enthusiasm.
• Hints of your personality: When managers bring on new staff, they want someone who they—and their team—will enjoy working with. A cover letter can give a better sense of who you are as a human being than a resume (which, in most cases, is a fairly impersonal document).

Additionally, if there’s anything in your resume that needs to be explained, like a gap in your work history, a well-constructed cover letter will do that.

Components of a good cover letter

All of this means that an ideal cover letter will:

• Be engaging without being overwritten or flowery
• Not sound generic, like it could have come from any other candidate
• Be enthusiastic about the position, but not gush so much that comes across as fake
• Highlight concrete details from your resume that showcase your strengths
• Explain anything that could be seen as a weakness

A good cover letter also won’t sound like it was written by an AI. If the person reading your letter suspects you didn’t actually write it, its ability to convey the qualities listed above collapses.

(The world is changing, and some hiring managers are coming to accept the idea that many candidates will use AI as a writing aid, but even they generally don’t want it to be obvious. Rightly or wrongly, a cover letter that was obviously machine-written can give the impression that you’re careless or lazy.)

I tried out several frequently recommended cover letter generator web apps. Some were entirely free, and others were “freemium” (meaning there’s a paid version that uses a more advanced AI model).

I used the free versions, my goal being to evaluate the output with typical (and not ideal) use. The majority of the users of any web app never fork over money if they can help it.

Testing Grammarly’s cover letter generator

Grammarly, which is mainly known for its proofreading software, has a cover letter generator. At the time of writing, it appears first when you search for “cover letter generator” in Google.

Most generators work in essentially the same way: you upload or paste your resume, answer a few questions, and the website spits out a letter. Grammarly’s app is no exception.

Using Grammarly to create a cover letter was very quick. It didn’t ask me many questions at all—just for the title and description of the role I was applying for.

I fed it a real ad from the Side Hustles job board (for a copy editing job) and my actual resume, which mainly lists my experience as a writer and editor.

I clicked “Create cover letter” in the interface shown above. Seconds later, Grammarly generated this:

Evaluating Grammarly’s output

Unfortunately, Grammarly didn’t do great.

I’ll start with the good: the cover letter above is clean and readable and would probably squeak past a hiring manager who wasn’t paying very close attention. Some AI-produced writing is very obvious and full of “tells,” such as excessive use of dashes and an odd obsession with words like “underscores.” This cover letter doesn’t have those flaws.

However, it has several other serious problems.

It undersold my experience

In the very first paragraph, the cover letter says I have “over three years of experience.” That’s technically true … with the emphasis being on technically. As I said at the top of this article, I have nearly ten years of experience as a writer and editor.

Grammarly (or, rather, the underlying AI model) failed to correctly interpret my resume.

This is a big deal. A cover letter that makes the candidate look less experienced than they are is quite literally worse than useless. It’s damaging, not helpful.

It’s way too long

By convention, cover letters usually consist of three paragraphs: one to express interest in the position, one to communicate why you’d be good at it, and one to politely sign off. This letter hits all of those notes, but it does it over the course of five paragraphs.

There’s no reason for it to be so long. A lot of this is pure fluff. The writing is very “adjective-y” (I don’t just have passion, I have a strong passion), and nouns almost always come in pairs (“collaboration and continuous improvement”).

All of this adds bulk. Way too much bulk.

In a world where a typical hiring manager spends only seconds reviewing each application, there’s a strong chance most of this simply won’t get read.

The experience cited is irrelevant

To its credit, AI has pulled several details from my resume. Unfortunately, they aren’t the right details.

There’s a sentence about my background in search engine optimization (SEO) editing. That would be great … except the job ad I used didn’t mention anything about the company’s content being SEO.

I have experience in my resume that would make me look like a great fit for this position, but that isn’t it.

Moreover, when the cover letter brags about my qualifications, its boasts are very jumbled. It jumps from teamwork to time management to collaboration in a way that makes the letter feel sloppy.

Overall, I give this cover letter a C-. It isn’t appalling, but it definitely isn’t good, either.

Testing Cover Letter Copilot

After wrapping up Grammarly’s somewhat disappointing test, I gave another generator, Cover Letter Copilot, a try.

Cover Letter Copilot is probably the most popular cover letter generator in the world right now; people often recommend it in job-hunting communities on Reddit.

It has a more elaborate interface than Grammarly’s tool, and allows you to enter much more information about what kind of letter you want. For instance, you can pick how formal you want your letter to be, with three possible options, “Professional,” “Modern,” and “Warm.”

I used the same job description and resume. Despite having a couple more steps, the process was still quick and easy (cover letter generators are meant to be time-saving tools, so they’re designed to be fast).

After I hit the button, Cover Letter Copilot thought for a moment and the AI model it uses (GPT) produced this. (I did this three times; this is the output for the “Professional” formality setting.)

Evaluating Cover Letter Copilot’s output

Unfortunately, this isn’t great, either.

I’ll start with the good: at least this is shorter than Grammarly’s output. It also gets the length of my career right—oh, wait, no, it doesn’t.

The AI wrote that I have “over a decade” of experience; that should be “nearly a decade.” Arguably, slightly overselling my experience is better than vastly underselling it, but it’s still far from ideal.

That also isn’t the only issue:

It’s overwritten (and identifiable as AI)

The slight “adjective-y” nature of Grammarly’s writing is even more apparent in Copilot’s output. “Sincere interest,” “meticulous editing,” “esteemed organization” … a little of that is OK, but too much really bogs down the letter and, as before, renders it much less likely to be read.

As an editor, I also know that this particular foible is common in AI writing. This means I can instantly tell that the letter above was produced using a generator.

Not all hiring managers are good judges of writing, and some will probably miss that. However, the skilled and experienced ones won’t. Such obvious AI use puts your application at risk of disqualification.

The experience still isn’t relevant

Like Grammarly, Copilot is bragging about the wrong things. For example, it cites my track record producing “medically reviewed content,” something that isn’t particularly relevant to the role I’m applying for, which entails editing content about home improvement.

There are potential ways to make that experience seem relevant, but this cover letter isn’t working very hard to do that, so it feels like I’ve misunderstood the job description or am just reusing a letter I wrote for another job. Neither is impressive.

I give this cover letter a C+. The shorter length means it’s less likely to raise eyebrows than Grammarly’s, and it didn’t make the disastrous mistake of underselling my experience, but it still isn’t much of an asset.

I was curious if these problems were inherent in AI—in other words, if the technology just wasn’t quite there yet—or if it was possible to coax an AI model into doing better.

I wrote up a set of instructions (known as a prompt) to plug into a model like Gemini or GPT. I told it to:

Avoid common AI-isms and writing mistakes

It goes without saying that I wanted my own tool to avoid the pitfalls I described above.

I stressed the importance of watching out for excessive adjective use, logical errors like bragging about irrelevant experience, and so on. My hope was that with more guidance, the AI would do better.

Guide the user through creating a cover letter

One of the problems I had with all of the generators I tried was that there was no back-and-forth dialogue. I thought a collaborative workshop would result in a better letter than the AI collecting my info and then generating everything without any further input.

Consequently, I told the AI to volunteer a few sentences to the user and ask if they were OK:

In a similar vein, I told it to point out any problems with the user’s resume, like an employment gap or a lack of relevant experience, and suggest ways to address them in the cover letter.

I didn’t want people to have to spend a lot of time haggling with the AI, so I told it to always suggest something and nudge people to just say “yes” if its solution was acceptable.

A good cover letter generator needs to be thorough and conscientious, but it also needs to be fast. People generally use these things because they’re firing off dozens of job applications and need to save time.

Testing my own generator

You can give the Side Hustles cover letter generator a try yourself. Follow that link or just scroll to the bottom of this article.

Here’s the output it returned when I took it for a spin (I did that several times; this is just one example).

Evaluating my own generator’s output

Obviously, I’m biased here, but not as biased as one might think. If my own prompt flopped, I’d cheerfully admit that I wasn’t able to use any of the AIs I tried to generate an acceptable letter.

Here’s my evaluation: The letter above isn’t perfect, and a skilled writer who spent real time on it would be able to do better, but this is also a pretty decent attempt.

It’s short and sweet (a good thing), lacks the obnoxiously “adjective-y” writing of the two I generated above, and is much more careful about only citing assets that are relevant to the job (editing speed and management experience) instead of ones that aren’t (knowledge of SEO).

I wouldn’t have submitted either of the first two letters I generated. I would submit this one. It wouldn’t be my first choice—I’d still rather write my own—but that’s because I’m a professional writer. Most people aren’t.

If I did submit this, my guess is that the majority of hiring managers wouldn’t be able to tell it was machine-written. A few savvy ones might, but if there are AI-esque patterns, they’re pretty subtle.

Testing the AI’s ability to spot red flags

I also played a tiny trick on my own generator. I uploaded my real resume, but changed my last job from “Editor” to “Meat Stacking Coordinator” at the “Lone Star Meat Shack” (which isn’t a real job or restaurant).

With my prompt, the AI caught that, pointed out that the irrelevant experience could be seen as a red flag, and suggested a way to address the issue. You can see that in the first paragraph: “while I recently spent time working directly in a hands-on labor role to gain a better understanding of the trades, I am eager to return to a full-time editing career.”

There isn’t a perfect way of taking a clearly irrelevant job and turning it into an asset, but this is about as good an attempt as it’s possible to make.

(When I performed the same test with Grammarly and Cover Letter Copilot, incidentally, neither caught the change to my resume or tried to address it in the letter.)

Paid vs. free AI models: is there a difference?

When you use any LLM like GPT, Gemini, or Claude, you can choose between an advanced (paid) model and an older (free) one. I tested my prompt with both.

GPT and Gemini’s paid models both generated noticeably better outputs and were less prone to getting confused in the conversation than the free ones. As noted, the same is true of products like Cover Letter Copilot—they have paid versions, which buy you access to a more sophisticated underlying model.

My goal with my prompt was to create something that will be usable without forcing people to spend money. Overall, I think I succeeded, although for the best results, you’d want to fork over cash.

If I’d stopped at the first two letters I created, then as an editor and hiring manager, I would have strongly advised against using an AI cover letter generator in 2026.

I’d have said the odds of the output getting clocked as AI—or of the generator writing something embarrassingly wrong—were just too high.

Fiddling with my own prompt changed my opinion. Now, I think submitting an AI cover letter isn’t necessarily a recipe for disaster.

However, I do think that if you do that, you shouldn’t use a premade web app. You don’t necessarily have to use my generator—although you’re welcome to!—but if you don’t, go directly to an AI model and have a quick conversation with it.

AI has advanced to the point that it’s capable of this task, but it needs guidance. If you let it off its leash, you can’t be sure you’ll get something usable.

The elephant in the room: is this ethical?

There are, of course, many people who have ethical or moral qualms with using AI for tasks like this.

Some people think it’s too dishonest or consider it plagiarism. Others take issue with the entire AI industry, citing environmental concerns about water and power usage.

This is something I haven’t tackled yet, and I’m not going to in this article. The topic is too big. It’s up to you to decide how you feel about AI-generated cover letters and whether you’re comfortable using them.

If you decide you can live with it, this editor’s verdict is that it’s possible to get away with it. But only if you’re smart about how you do it.

Don’t forget to edit your output

Even if you follow my advice, I recommend reading each letter you generate carefully and editing it yourself if necessary. Even a skillfully prompted AI model won’t get it right 100% of the time.

AI can be a time-saving tool. It can turn a 30-minute task into a 5-minute one. But it can’t cut that down to zero minutes.

If you want to give the cover letter generator I created a try, you can access it at that Gemini link. Just type an intro message like “hi”; the prompt will kick in and walk you through the process from there.

You can also take the full set of instructions I wrote below and paste them into any AI model that you like.

Do you have any comments about our cover letter generator? Did it create something you were able to use? Let us know on the Side Hustles subreddit.