How to Create Marketing Images With AI: Prompts and Edits
A repeatable three-step routine for turning a plain idea into a usable marketing image, then keeping every image after it consistent with the first.
AI Cover Letter Generator: How to Write One That Gets Read
A generated cover letter reads as generated in about four seconds. Here is the preparation that makes the difference, and the prompt that uses it.
ChatUp GuidesWork
Cover letters are the part of applying that everyone hates and most people do badly, which is precisely why doing them well is cheap leverage. A hiring manager reading forty applications can tell within a sentence whether a letter was written for this job or pasted into it.
AI makes the writing fast. It does not make the letter specific — that comes from what you feed it. So the work is front-loaded.
Do cover letters still matter?
Sometimes. At large companies with automated screening, often not much. At small companies, agencies, and anywhere a human reads applications, they matter a lot — and they matter most when you are a non-obvious fit: changing industries, returning after a gap, or lighter on one requirement than the posting asks.
If a letter is optional and your resume speaks for itself, skip it. If your resume needs explaining, the letter is where you explain it.
Prepare in three passes
1. Research the company properly
Prompt to try
Research [company] for a job application. What do they actually do and for whom, what's changed for them recently, what does their public communication suggest they care about, and what problems is a team like [department] likely dealing with right now? Give me sources. Flag anything you're unsure about — I don't want to reference something in a letter that turns out to be wrong.
That last instruction matters. A confidently wrong detail about a company is worse than no detail.
2. Mine the job description
Prompt to try
Here's the job description. Separate it into: hard requirements, soft preferences, and filler. Then tell me what the posting reveals about what's actually going wrong on the team — the problem they're hiring to solve, which is usually visible between the lines. What would they most want to hear addressed?
3. Establish your real evidence
Prompt to try
Here's my background: [paste resume]. For each hard requirement in that job description, find my strongest piece of evidence — a specific thing I did with a specific result. Where I have nothing, say so plainly rather than stretching something. I need to know my genuine gaps before I write.
The letter prompt
Now the generation step has something to work with.
Prompt to try
Write a cover letter for this role using only the evidence above. Rules: under 300 words, four paragraphs. Open with why this specific company and role, referencing something real from the research — no "I am writing to apply for." Second paragraph: the single strongest piece of evidence, with the result. Third: address the gap honestly and what compensates for it. Close with what I want to do there, not what I want from them. Plain, direct, professional. No enthusiasm adjectives, no "passionate", no "excited about the opportunity."
Then edit it. Read it aloud, cut anything you would not say in a room, and put back the one turn of phrase that sounds like you.
Reverse-engineering the letter
A stronger variant, especially when you are a marginal fit: write the letter the hiring manager wishes they would receive, then check how much of it is true about you.
Prompt to try
You're the hiring manager for this role. Describe the cover letter that would make you immediately want to interview someone — what it addresses, in what order, what it doesn't waste words on, and what would make you stop reading. Then compare that to my draft and tell me what's missing.
Common mistakes
- Restating the resume. They have it. The letter is for what the resume cannot say.
- Writing about yourself the whole way through. The best letters are mostly about their problem.
- Enthusiasm as a substitute for evidence. “I’m passionate about your mission” is a sentence anyone could write about any company.
- Length. Over one page is unread. Under 300 words is usually right.
- Untouched generated text. It has a texture, and people who read applications all day recognize it.
- The wrong company name. Reusing a letter without a careful find-and-replace ends more applications than any other single error.
Test it before sending
Prompt to try
Read this letter as a skeptical hiring manager with 40 applications and four minutes. Where do you stop reading, and why? Which sentences carry real information and which are filler? What would you ask me in an interview based on this — and is there anything here I couldn't defend?
Where this fits in ChatUp
AI Research for the company pass, since this is exactly the case where an answer from memory could be a year out of date. The Interview Coach assistant for the follow-through, because a letter that works produces an interview you then have to pass. File Assistant for working from your resume and the job description as documents. See also our guide on using an AI resume builder without sounding generic.
Frequently asked questions
Will they know it was written with AI?
They will know if you did not edit it. A letter built from real research about their company and specific evidence from your work does not read as generic, regardless of how the sentences were produced.
Should I use the same letter for multiple jobs?
The structure, yes. The opening paragraph and the evidence selection have to change per application, and those are the parts that get read.
How long should a cover letter be?
Two hundred and fifty to three hundred words. Three or four short paragraphs.
What if I do not meet all the requirements?
Address it in one sentence, without apology, and move to what compensates. Most postings list an ideal candidate who does not exist.
Preparation is the whole trick
The generation step takes ten seconds and has never been the bottleneck. Research the company, read the posting properly, be honest about your evidence, and the letter almost writes itself — which, at that point, is fine.
Try it in ChatUp
Turn this guide into a workflow.
Run the prompts above against the model that suits the task, keep the useful context across chats, and pick it back up on any device.
Try for Free