Killing his AI product after growing it to $250k ARR within a
year
After one successful exit, Ulric Musset launched an AI engineer called Marblism that quickly grew to $250k ARR. But then they had to pivot. Now they're rebuilding their offering — and their customer base.
Here's Ulric on how he's doing it. 👇
From exit to inception
Our fintech startup, Vauban, was bought by Carta when it was making $5M ARR. Straight away, my cofounder, Cyril Pluche, and I jumped into our next venture, Marblism.
It's a suite of AI employees that work for your business. This includes an Executive Assistant that filters emails and drafts responses, a Social Media Manager that's basically your personal ghostwriter, an SEO-focused Blog Writer that constantly pushes new articles to your website, and a Sales Assistant that generates leads while you sleep.
We quickly hit $250k ARR.
Raising $4M
After exiting our last startup, we knew we wanted to create something that would make it simple for entrepreneurs to start and grow their businesses.
We launched in January 2024 with an AI Engineer capable of building full-stack applications in minutes — something no one else was offering at the time. Being first to market with this capability gave us confidence in our direction.
As far as our stack, we practice what we preach – our platform runs on the same tech we originally built to generate apps: React frontend, Node.js backend, all hosted on AWS.
Under the hood, we're running customized LLMs that we've fine-tuned for specific tasks like emails, agentic behaviors and content creation.
Three months after our initial launch, we raised $4M.
Surprisingly, investors seemed more impressed by our YC badge than by our previous successful exit. The AI industry has reshuffled priorities — investors now emphasize deep technical expertise, explosive user traction, or that YC stamp of approval.
A Hard pivot
A year into our journey, we pivoted dramatically. Initially, we had assumed developers would be our main audience, but there was a plot twist: Business owners without coding skills became our primary users instead.
This unexpected audience helped us quickly reach 100,000 users and achieve $250K in ARR sooner than we anticipated. However, we encountered two significant problems: Most customers wanted to build AI applications or AI agents to assist with their businesses. And, more importantly, resolving minor bugs proved extremely challenging for users without coding experience.
Unfortunately, these were problems we couldn't solve with our existing approach. We had to pivot.
It wasn’t our first pivot — we had pivoted in our previous venture as well. When you're facing such a tough decision, you can feel in your gut that something isn’t right and that it's time to change course. After a week of reflection, we took action.
In just three months, we transformed our AI full-stack engineer product into a comprehensive team of specialized AI employees. Having the financial runway from raising capital gave us the flexibility to quickly course-correct and evolve without facing existential risk.
We relaunched a week ago and we've got a long way to go to get back to our peak revenue of $250k.
Marketing is not rocket science
Marketing a product isn't rocket science – you try everything until something works: SEO, Twitter, cold outreach, paid keywords, Meta ads, etc..
For us, what's working is word-of-mouth, SEO, and Meta ads.
Not all products have equal word-of-mouth potential. Your product's outcome needs to be inherently shareable, or using it should make people feel smart or cutting-edge when they mention it to their friends.
Overall, I'd say the main factor in our success was timing. Timing is everything. We launched right as AI was exploding, creating natural momentum for our product and helping us quickly attract over 100,000 users.
People are surprisingly willing to try a lot of AI tools right now. Getting paid users isn't the challenge; the real test is building something good enough that people stick around. Acquisition is just the beginning — retention is where the war is won or lost.
Educate your users
If I could hit reset, I'd invest heavily in user education from the start. Helping customers understand AI's capabilities and limitations is crucial — especially when their expectations have been shaped by unrealistic tech headlines.
Something a lot of AI tech founders take for granted is that their users will know how to prompt AI perfectly.
The reality is very different. To address this, we've designed our AI employees to proactively "reverse-prompt" users, meaning they reach out to users only when they need approval or additional information. This simple adjustment has really improved user experience.
Find the right pricing model
At first, we planned to charge per AI Employee, which made sense since it's like paying a small salary. But we realized that new concepts need to be simple, so we switched to a package with all AI employees included.
We relaunched two weeks ago and already have 100 new customers. Our margins are good because we only pay for computing resources and AI improvements. We've built systems that use fewer tokens while maintaining quality, which keeps our costs down as we grow. It's our secret sauce!
Don't marry your vision
Don't marry your initial vision. Be ready to pivot when users tell you (directly or through their behavior) that you've built the wrong thing. Our transformation from AI engineer to AI employees came from watching how people actually used our product versus how we imagined they would.
And each story is unique, so don’t follow advice blindly, craft your own path. Follow your gut.
What's next?
We want to make AI employees as commonplace as email for small businesses. Our target is hitting $1M ARR by year-end by expanding our AI employees roster.
Our biggest ongoing challenge is to manage expectations in a hype-filled industry. We're committed to honest communication about what our AI can and can't do — avoiding the overpromise-underdeliver trap that leaves users frustrated.
We also want to pioneer the next evolution in AI by developing collaborative multi-agent teams. Instead of relying on a single AI to handle all tasks, we’re building systems where specialized agents work together, communicate, and delegate tasks to one another, just like a real team.
This unlocks a more powerful and natural way for businesses to operate.
You can follow along on X and LinkedIn. And check out Marblism.
Leave a Comment
Bold move — killing a $250K ARR product shows he's chasing vision over comfort.
Or churn was too high
Got to know when to pivot.
This is a fascinating insight into the reality of building an AI startup. Scaling to $250K ARR in a year is no small feat, but the decision to pivot highlights an important truth—market fit is everything. AI is powerful, but if the end user struggles with implementation, even a great product can hit a ceiling. Excited to see how this new direction plays out!
Really appreciate the honesty here. Most founders gloss over the hard pivots, but this breakdown shows how much of the real work happens after the initial traction. The “don’t marry your vision” point hit hard, user behaviour > founder assumptions every time. Curious to see how the multi-agent approach plays out.
Thanks for sharing such an honest journey, Ulric! Your pivot story really resonates. At RaftLabs, we've also seen how critical it is to identify your true audience quickly. The idea of 'reverse-prompting' for user education is brilliant — a smart way to bridge the AI knowledge gap for non-technical users. Best of luck hitting that $1M ARR target!
A few things that really stood out:
Just because something is working doesn’t mean it’s worth continuing.
The product was growing, the money was coming in, but it didn’t feel right. That’s a side of entrepreneurship that doesn’t get talked about enough.
Saying “no” is sometimes the smartest move.
Walking away from $250K ARR takes guts, especially when the numbers look great on the outside. But if it’s draining you and you’re not excited to keep building, it’s a dead end in disguise.
Speed hides cracks.
When something grows fast, it’s easy to ignore the parts that don’t feel sustainable: the support load, the infrastructure mess, the “I’m not sure I want to be the person who runs this” feeling.
There’s more than one definition of success.
It’s refreshing to see someone define success not just by revenue, but by how a project fits their life, energy, and long-term direction.
I’d love to hear:
– What helped you make the final call to shut it down?
– How are you approaching your next project differently?
Scaling to $250K ARR in a year is impressive, but the pivot really highlights how crucial market fit is. Even the best AI tools can stall if users struggle with adoption. Curious to see what’s next!
By the way, if you're into mobile gaming, check out our latest release: Car Parking Multiplayer 2 Mod APK — we’ve added some exciting features!
Honestly, this is brutal but super real. People don't realize how fast building something and burning out from it can happen at the same time. I respect the hell out of pulling the plug when it doesn't feel right anymore — that's not weakness, that's control. Everyone loves to talk about scaling, ARR, "exit velocity," whatever, but nobody ever talks about the personal cost when your own project turns into a prison.
$250K ARR sounds sexy until you're living in a day-to-day grind you resent. I'd rather build something I'm obsessed with for $50K ARR than hate something that "succeeded" on paper.
Appreciate you sharing this — most people would just quietly fade out and pretend they were still "crushing it."
Really insightful journey — especially the pivot and focus on user education. It’s refreshing to see such transparency about the challenges and mindset shifts along the way. Timing really is everything, but pairing it with adaptability is what makes it work. I’ve gone through similar transitions with my own project in the IPTV space: lite4kiptv dot net. Thanks for sharing — lots to learn here!
This story proves that even with impressive growth, not every product is sustainable in the long run. The founder’s honesty and timely decision to shut down is truly inspiring. In the fast-paced tech world, clarity and adaptability are key.
Wow, that's a wild ride! $250k ARR in a year and then a pivot? Gotta admire the hustle and adaptability. Good <a href="https://crazycattle.pro/" target="_blank">luck with the relaunch!</a>
Bold story. What stood out most is how quick they were to pivot once user behavior didn’t match expectations.
Makes me think how many branding or positioning choices out there get killed too early just because they don’t fit “the usual mold”?
Sometimes it’s not about what’s right or wrong just what’s too early or unfamiliar.
Installing Insta Pro 2 is simple, but since it is not available on the Google Play Store, it requires a manual installation process. Here are the general steps:
My marriage was restored and my husband came back to me he apologized for all the wrongs he did and promise never to do it again. A big thanks to this wonderful psychic for bringing my husband back to me.. I never really believed in magic spells or anything spiritual but a trusted friend opened my eyes to the truth about life. My marriage was heading to divorce a few months ago. I was so confused and devastated with no clue or help on how to prevent it, till I was introduced to Dr. Excellent that did a love spell and broke every spiritual distraction from my marriage. A day later my husband started showing me love and care even better than it used to be, he’s ready to talk things through and find ways for us to stay happy. It’s such a miracle that my marriage can be saved so quickly without stress. You can also contact him for help. Here his contact. Call/WhatsApp him at: +2348084273514 "Or email him at: Excellentspellcaster @gmail. com
Sometimes the hardest part isn’t starting — it’s walking away.
Respect for sharing the shutdown openly.
AI products can scale fast, but sustaining them without alignment or excitement is a different game entirely.
I love reading stories like this, not just about the ups and downs, but also about the painful twists and turns. It’s really cool that the team didn’t get stuck on their original vision, but were able to rethink their target audience and take a new step. It’s honest, candid, and truly inspiring.
I love the idea behind your pivot: "Your own AI workforce" is a much stronger value proposition than "an AI that writes software". Not so sure the bold move is to make the pivot. It is very likely difficult to compete within the "AI engineering" space because it attracts so much attention, doesn't it? It seems like a smart move that you moved towards low-tech businesses with a need for AI (much bigger market).
Your "AI workforce" idea is very inspiring. I'm currently trying to develop automation software for small sales agents in the insurance space and your value proposition is probably also very attractive for that market segment (more attractive than "automate your processes" - even though it's the same thing on a fundamental level).
Good luck !! <3
The tech industry with "no/low code", "AI coders", and "vibe" is making promises to the "citizen developer" that it knows can't be kept. Home Depot offers to help DIY'ers replace a toilet seat (to avoid paying a plumber) not add a room addition to your house.
Thanks for the post! I could learn a lot from it. You said if you could hit reset, you would have invested heavily on user education. Do you think that if you did educate the users before the pivot, you wouldn't have to dramatically pivot?
Also, I'd like to learn from you how to educate users to align with the service vision and purpose. Some customers just demand their needs to the maker, instead of trying to fully understand about the service they're using.
insightfull....
Insanely honest and valuable post. Growing an AI product to $250K ARR only to shut it down takes guts — but it also shows deep self-awareness. Love the breakdown of what didn’t work, especially the insight around founder-market fit and energy levels. A must-read for anyone chasing 'easy' wins in AI.