Resign AI - Resignation letters
An AI-powered web application for creating professional resignation letters. Input personal and job details, customise tone and formatting preferences, and generate tailored letters. Featured credit-based pricing with multiple customisation options.
The Origin Story
After building Joblet (my AI cover letter generator), I had a slightly cynical realisation: if people need help getting jobs, they probably also need help quitting them. So I built ResignAI, the exact same concept but in reverse. Instead of helping people write letters to get hired, it helps them write letters to quit. The irony was not lost on me.
This was basically Joblet 2.0. Same tech stack, same AI integration approach, same credit system, same everything. Just different prompts and a different target audience. I wanted to see if I could replicate the success (and lessons learned) from Joblet but target a different pain point in the employment lifecycle.
Spoiler: it worked just as well technically, and met the exact same fate. The market for AI-powered resignation letter generators is, unsurprisingly, also overcrowded. Who knew? (I should have, honestly.)
After running it for a while, I took it offline for the same reasons as Joblet. The learning objective was achieved, the market was saturated, and I didn't see a compelling path to differentiation. This is that story.
What ResignAI Actually Did
ResignAI was a full-featured web application that generated professional resignation letters using AI. The workflow was nearly identical to Joblet:
- Enter your details, full name, job title, company name, manager's name, and last working day
- Customise your letter, select tone (casual/formal/professional), length (short/medium/long), creativity level, language variant, and date format
- Get your letter, AI generates a polished, professional resignation letter in seconds
The optional content sections let users really personalise their letters:
- Reason for leaving, with specific details about why you're moving on
- Offer of transition assistance, showing you're leaving responsibly
- Expressions of gratitude, acknowledging positive experiences
- Mention of positive experiences, specific highlights from your time there
- Address placeholders, for proper business letter formatting
The AI would weave all of this together into a professional resignation letter that maintained good relationships while clearly communicating your intentions. Because burning bridges is rarely a good career move, even when you really want to.
The Technical Side
Since this was essentially Joblet with different prompts, the tech stack was identical:
The Stack
- Laravel 12, latest and greatest
- Livewire 3, reactive components without JavaScript
- Tailwind 4, fast, clean UI development
- Anthropic API, for letter generation and blog content
- Open AI, for blog post images
- Stripe with Laravel Cashier, payment processing and subscription management
AI Prompt Engineering
The interesting challenge here was getting the tone right. Resignation letters are tricky, you want to be professional and positive even if you're leaving because you hate the place. The AI needed to:
- Stay professional regardless of the user's actual feelings
- Express gratitude authentically without sounding fake
- Clearly state intentions while maintaining warmth
- Offer help with transition without over-committing
- Handle sensitive reasons for leaving tactfully
The creativity levels (Precise, Balanced, Dynamic, Creative) adjusted how "safe" versus "personable" the letter felt. Precise was very corporate and formal. Creative was warmer and more human while still being professional.
Credit System & Payments
Same system as Joblet, powered by Stripe and Laravel Cashier:
- 2 free credits for new accounts
- Additional credits via one-time purchases
- Three tiers: £3.49 for 5 credits, £7.49 for 15 credits
- Custom webhook handler for automatically assigning credits after successful payments
Pricing was slightly lower than Joblet because resignation letters felt like a one-time need rather than something people would use repeatedly. Most people only quit a job once every few years.
Automated Blog Content
Just like Joblet, I built an automated content marketing system:
- AI-generated blog posts based on keywords and prompts
- AI-generated images for each post using DALL-E
- Queued to publish once per week over a 6-month period
- Topics like "how to resign professionally", "career transition advice", "maintaining workplace relationships"
Six months of content, fully automated. Set it and forget it. The blog covered everything from "how much notice to give" to "what to say in your exit interview" to "negotiating your last day".
User Experience
Built with Livewire for full reactivity. Fill in your details? The preview updates. Change the tone? See what that means. Adjust creativity? Get instant feedback.
Generated letters were stored in your account with full CRUD operations. View, edit, download as PDF, delete. Everything you'd expect from a document generation tool.
What I Learned
1. The Pattern Is Repeatable
Building ResignAI proved that the Joblet architecture and approach worked for different use cases. Same tech stack, different domain, same quality results. This validated my AI integration patterns and made me confident I could build similar tools quickly.
The entire build took about a week because I was essentially copying and modifying Joblet. Most of the time was spent on prompt engineering and tweaking the UI copy to be resignation-focused rather than job-application-focused.
2. Market Validation Still Matters
Even though I knew the AI letter generator market was crowded from Joblet, I built ResignAI anyway. Classic mistake. Just because the tech stack is proven doesn't mean the market opportunity exists.
There are fewer resignation letter generators than cover letter generators, but the market is still saturated. And importantly, people resign far less frequently than they apply for jobs, so the total addressable market is smaller.
3. Tone Is Everything for Sensitive Topics
Resignation letters are more emotionally charged than cover letters. People are often leaving because they're unhappy, underpaid, or burned out. But they still need to maintain professionalism and avoid burning bridges.
Getting the AI to generate letters that felt genuine while staying positive required careful prompt engineering. Too formal felt cold. Too casual felt unprofessional. Too enthusiastic felt fake when someone was clearly unhappy.
The balance took iteration, but I eventually landed on prompts that produced letters that felt authentically professional regardless of the user's actual feelings about their job.
4. One-Time Use Limits Growth
Unlike Joblet where someone might generate multiple cover letters for different applications, most people only need one resignation letter per job. Some might tweak it a few times, but it's fundamentally a one-time-per-resignation tool.
This limits the credit consumption model. Most users would sign up, use their 2 free credits to generate and tweak a letter, then never come back. The economics don't work for sustained growth without either raising prices significantly or pivoting to a different model.
Technical Challenges
Most of the technical challenges were solved in Joblet, so ResignAI was relatively smooth:
Challenge 1: Privacy Concerns
Resignation plans are sensitive. Users were understandably nervous about uploading employment details to a third-party service. I emphasised encryption and privacy throughout the site, but it was still a barrier for some users.
Solution: Added clear privacy messaging, SSL everywhere, no data sharing promises, and the ability to immediately delete generated letters. I also kept all processing in-memory where possible and didn't log sensitive details.
Challenge 2: Handling Negative Situations
Some users wanted to express frustration or negative reasons for leaving. The AI needed to translate those into professional language without losing authenticity or making the user sound bitter.
Solution: Refined prompts to acknowledge difficult situations while reframing them professionally. For example, "toxic workplace" became "seeking a better cultural fit" and "terrible manager" became "looking for different leadership style".
Challenge 3: Legal Concerns
Some users asked if the generated letters were "legally valid" or if they needed to add specific clauses. I'm not a lawyer and didn't want any liability.
Solution: Added clear disclaimers that the tool generates template letters and users should consult HR or legal counsel for specific situations. Also included standard "this is not legal advice" language throughout.
Why I Shut It Down
Same story as Joblet:
- Market saturation, too many similar tools already exist
- Low differentiation, nothing made ResignAI uniquely better
- Limited repeat usage, one-time-per-job-change model limits growth
- Mission accomplished, proved the pattern works for different domains
The goal wasn't to build a SaaS empire. It was to validate that the AI integration patterns from Joblet could be applied to other document generation use cases. Mission accomplished.
I gained additional experience with:
- Reusing and adapting existing AI architectures for new domains
- Prompt engineering for sensitive/emotional topics
- Building tools that handle private/confidential information
- Refining credit systems and payment flows based on Joblet learnings
All of that knowledge has been valuable in other projects where AI-generated content actually provides meaningful differentiation.
What I'd Do Differently
If I were building this again (which I won't, but hypothetically):
-
Target B2B instead of B2C. HR departments, career coaches, and outplacement firms might have been better customers than individuals
-
Bundle with other career transition tools. Resignation letter alone isn't enough. Add exit interview prep, references management, LinkedIn optimisation, etc.
-
Skip it entirely. Honestly, the lesson from Joblet should have been "don't build another AI letter generator" not "build the opposite kind"
-
Focus on the emotional support angle. The real value wasn't the letter, it was reducing stress during a difficult transition. Should have leaned into that more
-
Make it free and find another monetisation model. Charging per letter for a one-time-use tool creates friction. Maybe monetise through job boards, recruiting agencies, or career coaching partnerships instead
The Real Takeaway
ResignAI taught me that sometimes the best lesson is the one you already learned. I built Joblet and learned the AI letter generator market was too crowded. Then I built ResignAI anyway and learned... the AI letter generator market is too crowded.
But that's okay. The second time around was faster, easier, and validated that my technical approach was solid. Sometimes you need to make the same mistake twice to really internalise the lesson.
The real value was proving I could quickly spin up AI-powered tools in different domains. That skill has been way more valuable than either Joblet or ResignAI as standalone products.
Plus, now I have experience with both sides of the employment lifecycle. Getting hired and quitting. The circle is complete.
Conclusion
ResignAI was a successful technical project that taught me to trust my initial instincts about market saturation. The tech worked great, the user experience was smooth, the AI generated quality letters. But none of that matters if the market is already solved.
Sometimes you build things to learn, not to launch. Both Joblet and ResignAI fall into that category. They were excellent learning projects that taught me valuable skills I've applied elsewhere.
And hey, at least I can confidently say I've built AI tools to help people with both the best and worst days of their careers. That's something.
Would I Recommend Building This?
Nope. Learn from my mistake. If you want to practice AI integration, build something in a less saturated market. Or build it as a pure learning project and don't bother launching.