Robin Labrot

The decks you need to build successful (and cool) products ❘ Founder @ListenUp!

Most tools cover ~80% of what your team needs.
That’s because SaaS vendors will always optimize for the average Joe.

(I know, I’ve been on the other side of the fence, building products that reached +$25M ARR)

Nothing wrong with that. But your business runs on specifics.

So I’m building something to close that gap.

SparkUp! → a no-code way to bring AI and automation directly into the tools your team already uses.

It lets Sales & Support Ops add custom buttons or mini-UIs inside any SaaS app to trigger their Zapier, Make, n8n, or Relay workflows right where work happens.

The result: your team boosts productivity by triggering the right actions at the right time, inside their existing habits.

Here on LinkedIn, I share how I’m duct-taping the internet one Spark at a time to help you make your stack smarter, simplify tool chaos, and bridge the gap between what your tools do and what your team actually needs.

I’m doing all this as a bootstrapped founder, sharing with radical transparency what the journey is really like.


Want to get in touch?
→ DM me here
→ or email robin@sparkup.design

Head Of Product

I’m happy to share that I’m starting a new position as Founder at TalkToHumans!

TalkToHumans: turn your LinkedIn inbox into a Sales machine ❘ ex-lemlist

Head of Product @ Digilityx - We transform ideas into impactful products! Co-founder & Product Marketing @ Blockchainr - Building the next Craigslist for NFTs

I've been building products for 10 years. From sports Apps in a startup environment, building a Product Portfolio in the healthcare industry, growing my own SaaS product to building a billion-dollar e-commerce from scratch, I've seen a lot of different product approaches.

The most important thing I’ve learned about myself : I love people, business & technology. 

This drove me to my current position as a Lead Product Manager. Our team is currently working on rethinking an e-commerce platform from the ground up!

Growing up and living in different countries taught me the following about my skills:

I have a great gut instinct but also firmly believe in data : I base my decisions on facts and observations
I am passionate and driven. In the words of Seth Godin : "persistence isn’t having the same tactics over and over, it's having the same goal over and over"
I solve problems, lead and drive multicultural teams to give the best out of themselves
I speak tech and exec
I’m a strong believer in collective intelligence : I’ve learned that two brains are practically always better than one
I get people engaged : half of my job is getting different entities to work together
I am calm and focused in crises
I try to see every day as an opportunity to learn something new. If I don’t know it, I’ll learn it!

📢 My wife says that if she left me in an empty room, I would make friends with the door, so don’t be afraid to reach out and have a chat.

Cedric Blanchard

If you spend an hour with Jordane Giuly, 4 things will strike you:

𝟭/ 𝗛𝗶𝘀 𝗯𝗿𝘂𝘁𝗮𝗹 𝗵𝗼𝗻𝗲𝘀𝘁𝘆
He'll tell you he spent 18 months building a product nobody wanted. That at Spendesk, whenever there was a problem, the solution was "let's hire someone". That at Defacto , they lent money to a company because they didn't detect a corporate bond issuance in their transactions.

Zero bullshit. Just the facts.

𝟮/ 𝗛𝗶𝘀 𝗰𝗹𝗮𝗿𝗶𝘁𝘆 𝗼𝗻 𝘄𝗵𝗮𝘁 𝗵𝗲 𝗿𝗲𝗳𝘂𝘀𝗲𝘀 𝘁𝗼 𝗿𝗲𝗽𝗲𝗮𝘁
Most founders talk about what they want to build.
Jordane talks first about what he refuses to reproduce.

"We didn't want to create a company where you have to hire to scale."
"I didn't want staff meetings to revolve around org structure instead of product."

That kind of self-awareness doesn't come from an MBA. It comes from walking away from a unicorn mid-flight because the challenge had changed nature.

𝟯/ 𝗛𝗶𝘀 𝗼𝗯𝘀𝗲𝘀𝘀𝗶𝗼𝗻 𝘄𝗶𝘁𝗵 𝗲𝗱𝗴𝗲 𝗰𝗮𝘀𝗲𝘀
When Jordane talks product, he doesn't sell you a vision. He talks about the long tail.

How they underestimated that one bank account could contain multiple businesses.
How "capital increase" in a transaction doesn't always mean fundraising.
How in FinTech, building a product is like building a watch.

A PM who understood that the devil is in the details and that execution beats vision.

𝟰/ 𝗛𝗶𝘀 𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝘁𝗼 𝗶𝗺𝗽𝗼𝘀𝗲 𝗮𝗰𝗰𝗼𝘂𝗻𝘁𝗮𝗯𝗶𝗹𝗶𝘁𝘆
At eFounders, Jordane had two non-negotiable weekly meetings: product committee and go-to-market.

At Defacto, they're 30 people and have moved over €1 billion in credit.

The secret? "It's easier to be accountable to someone else than to yourself."

That discipline transformed a Stanford engineer who found machine learning "not that interesting" into the CEO of one of Europe's most promising FinTechs.

---

People often say the best product people have something extra.

After talking with Jordane:
- He's one of them.
- His thing is understanding that the perfect product doesn't exist. Just honest iterations.

---

I had the chance to record 55 minutes with Jordane about his journey from Spendesk to Defacto , his mistakes, and why most founders build in secret only to launch to silence.

Full episode link in the comments 👇

Learn from French builders, apply it with AI, ship better products | Host @Le French Product 🎙️

Everyone's asking: how do we adopt AI faster?

Alan's answer: you already missed the first step.

I went into this episode thinking I was going to talk about AI.

Turned out we mostly talked about culture.

Thomas Rolf is the CPO of Alan - one of France's most watched unicorns. And when I asked him why Alan adopts AI so much faster than everyone else, he didn't talk about their tech stack or their budget or some genius hiring decision.

He talked about transparency. Autonomy. What happens when you trust people enough to give them full context and let them act on it.

"The culture came first," he said. "The AI just landed on fertile ground."

I've been thinking about that ever since.

Because I've seen a lot of companies try to bolt AI onto a culture that wasn't ready for it. And I've seen how that goes.

Alan didn't plan to be AI-ready. They just built the right culture, and the AI fit.

Full conversation in the comments if you want the whole picture 👇

Most vibecoded apps look like they were built by someone who's never used one.

Every feature works. Nothing crashes. The AI nailed the logic.

But you open the app and... you feel nothing.

No personality. No warmth. No reason to come back.

I keep getting back to something Nesrine Changuel, PhD told me on Le French Product.

The number one feature used on Slack isn't messaging. It's custom emojis.
Let that sink in.
The core product is chat. But what people actually love (what makes Slack feel like their space) is a tiny surface detail that has zero to do with the core job-to-be-done.

Nesrine calls it surface delight. The 10% of effort you put into the things people see, touch, and feel, not the things that show up on your PRD.

And this hits different right now.

Because vibecoding is making it insanely easy to ship functional products. I'm doing it myself every week. You can go from idea to working prototype in hours.

But functional isn't memorable. Functional is table stakes.

The gap between a vibecoded side project and a product people actually want to use? It's not in the logic. It's in the surface. The micro-interactions. The personality. The thing that makes someone smile before they even complete their first task.

So now, every time I vibecode something, I force myself to ask: where's the custom emoji moment?

Where's the 10% that turns "it works" into "I love this"?

If you're shipping fast with AI  (and you should be) don't skip the delight.

That's the moat now.

What's a product that nailed surface delight for you? 👇

A few months ago I called OpenClaw a "hard pass." An experimental open-source bot, a viral moment with no business model, a tool for hackers not professionals.

I meant it.

Then I watched Aakash Gupta's episode with Naman Pandey and had to sit with some real discomfort.

The use cases weren't dev demos. They were built for PMs. A Slack channel that becomes a queryable knowledge base from your PRDs and wikis. An automated 9am standup brief that scans overnight activity and posts a summary before you open your laptop. A competitive intel agent that monitors competitor sites around the clock and drops SWAT briefs into a private channel.

That last one stopped me. Catching a competitor change at 3am that disappears by morning. Persistent memory. Never sleeps.

No other tool does that.

But the security anxiety is still real. During the episode, the host accidentally had the bot message everyone on his WhatsApp with auth codes. A live security audit revealed unrestricted file system access and a disabled firewall. That's not a footnote. That's the whole conversation.

What surprised me: the episode also showed a path. Ask the bot to audit its own vulnerabilities. Run it in a dedicated environment. Set hard limits.

My previous dismissal was too clean. The value is there. The risk is there. Both are real.

I'm not doing a full reversal. But I'm willing to say I might be leaving real value on the table, and that's a position worth sitting with.

What do you think?

Product Managers have a Breakup problem.

And it’s limiting them.

Stay with me.

When you go through a breakup, it leaves a scar.
You don’t love the same way after that.
- You hesitate.
- You overanalyze.
- You protect yourself.
Now apply that to building products.

If you’ve been in tech 10+ years, you’ve accumulated scars:
 • “Can't do this.”
 • “That won’t scale.”
 • “Engineering won’t accept this.”
 • “We tried this in 2018.”
 • “This takes 6 months minimum cause you have to plug the thing into the other thingy, then make sure this communicates with that and...”

You think it’s experience.
But often… it’s just trauma from a different tech era.
_____________________________________________
Meanwhile, a 19-year-old opens AI dev tools and thinks:

“Why not?”
 
No legacy PTSD.
No Jira flashbacks.
No architecture heartbreak.
Just curiosity.

And sometimes? They ship faster than your entire squad.

The real constraint in 2026 isn’t tech.
It’s cognitive debt.

We are limited by problems that may no longer exist.

Hard question for product leaders:
Are you protecting the company…
Or protecting yourself from reliving past failure?
___________________________________________
If you want to fight the Breakup Effect:
1️⃣ Run a “naive rebuild” session.
2️⃣ Take a step back on a regular basis. I use a physical timer to trigger this consciously.
3️⃣ Prototype before debating.

Because sometimes experience isn’t wisdom.

It’s just emotional residue.

The best builders today aren’t smarter.
They’re less scarred.

And that changes everything.

Focus is the number one problem when building with AI.

Not speed. Not technical skill. Not funding. Focus.

Here's the trap I see playing out everywhere. You start with one clear problem to solve. You fire up your AI tools. Within an hour, the AI suggests a related feature. Then another. You keep saying yes because each suggestion sounds smart in isolation.

Three weeks later you have a massive product with a million features. Hard to maintain. Impossible to validate. And nobody can explain what it actually does.

AI is incredible at generating ideas. That's exactly what makes it dangerous for builders without discipline.

Every successful company started by doing one thing exceptionally well. Salesforce didn't launch as a CRM suite. It solved one narrow problem, validated it, then scaled. The same logic applies today — maybe even more so, because AI makes it so easy to skip that step.

The 90% of building that AI handles for you means nothing if you pointed it in 12 directions at once.

Do one thing really, really well. Validate it. Then scale.

Focus is the number one problem when building with AI.

Not speed. Not technical skill. Not funding. Focus.

Here's the trap I see playing out everywhere. You start with one clear problem to solve. Yo

Quit wasting Claude tokens on sloppy prompts. 💸

I spent my weekend trying something Xinran Ma mentioned on Aakash Gupta’s latest podcast: use Gemini to write the prompt, then hand the clean spec to Claude.

If you’re a PM or builder, you know the token pain. But the bigger point: “build me an app” isn’t design.

Design with AI is workflow + constraints + clear intent.

Two workflows worth stealing:

1. The token-saver PRD workflow 
Don’t open Claude Artifacts, v0, or Lovable with a blank page. That’s how you burn tokens and go nowhere.
• Get clarity elsewhere: use a cheap/free model (Gemini/ChatGPT) as your “clarity engine.”
• Force constraints: make it ask you questions like target user, platform, and the one core flow. Tip: skip login screens. Start at the value.
• Write a markdown spec: front-end focused, strict, and copy-pasteable.
• Build in the expensive model: paste the spec into your coding model and generate the UI.

2. The divergent ideation workflow 
AI drifts toward the average. If you want fresh concepts, push it wide.
• Start with reality: screenshot an existing feature.
• Go wide: drop it into Google Stitch and turn on YOLO mode to get 15+ different directions.
• Prototype fast: export the winner into Google AI Studio and make it interactive.

Bottom line: anyone can get AI to spit out something. It will also make things up and default to generic.

The PMs who win won’t just do the same job faster. They’ll build repeatable workflows that change how the work gets done.
—
Want Workflow #1? 
 
I pulled the exact logic from the episode and turned it into a copy-paste system prompt. Simple, pragmatic, tested.

Paste it into a Custom GPT, a Gemini Gem, or a Claude Project to turn that model into a “clarity engine” that interviews you and writes PRDs.

Comment “CLARITY”, connect with me, and I’ll DM you the raw markdown file so you can use it right away. 👇

taplio

The Battle

Robin Labrot

Cedric Blanchard

Followers

Posts

Engagements

Likes

Performance ScoreRobin Labrot's LinkedIn Performance

Performance ScoreCedric Blanchard's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Robin Labrot

Cedric Blanchard

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Robin Labrot's LinkedIn Performance

Performance ScoreCedric Blanchard's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreRobin Labrot's LinkedIn Performance