Entrepreneur, Start-ups lover, Growth Thinker

Obsessed with hard / valuable / fun problems

Robert Goncear

Our users kept referring people.

Knowing we had no way to reward them.

For months.

New users would sign up and tell us "a friend recommended Baseloop."

We loved hearing it.

But it also meant someone out there was bringing us customers, 

and all we could say back was "thank you."

No referral link to share.
No credits to offer the person they referred.
No payout for the one who sent them.

They did it anyway. Because the product worked and they wanted others to have it too.

That's the kind of thing you don't take for granted.

Today we finally made it official.

Baseloop's referral program is live for all users, powered by Cello.

Here's what it means for you:

→ You get a personal referral link inside Baseloop
→ When someone you referred becomes a customer, they get 2,000 free credits
→ And you earn $50

No hoops. No tiers. You share, they sign up, you get paid.

If you've already been telling people about Baseloop, now there's a link for that.

Shoutout to Ana Marie Nila Hernandez, FIA, Jesse Singer, and Glenn Perez for getting this integrated.

And big thanks to Cello for powering it.

Our AE makes 400 cold calls a day.

14% connect. That's ~56 live conversations.

After every connected call, there's a checklist no one talks about:

→ Log the disposition
→ Update the CRM
→ Pause the outbound sequence
→ Write a follow-up based on what was said

That's 10 minutes of admin per call. Multiply that by 56.

Vincent (our AE) flagged it pretty fast:

At this volume, he couldn't follow up on everything.

Conversations were slipping through.

Not because the calls went badly, but because there wasn't enough time between them.

So we automated all of it 👇

Every prospect starts in a table where we enrich them with verified phone numbers from FullEnrich and BetterContact.

From there, they get pushed into lemlist, which runs the full sequence: emails, follow-ups, and call steps.

Salesfinity dials those call steps at scale.

When someone picks up, this is where it gets interesting:

→ The transcript and disposition flow back into the table automatically
→ Each call gets scored: interested or not
→ Interested? The lemlist sequence pauses, the transcript gets analyzed, and a personalized follow-up goes out
→ Not interested? The sequence continues or they get removed

All of this runs in Baseloop.

Zero admin between calls. Vincent van der Plas just dials.

Cold calling at scale only works if your post-call workflow can keep up. Now ours does.

Our AE has 40+ follow-up tasks every Monday.

Prospects who said: not now.
Reach out in a few weeks.

Good conversations. Real interest. Just bad timing.

Here's what clearing that list actually costs:

→ Open a task
→ Open HubSpot
→ Find the last email
→ Scroll through call notes
→ Remember what was said
→ Write something that doesn't sound like a template

5 minutes each. 40+ tasks. Half a day gone before a single email is sent.

We decided that's AI's job. Not Vincent's.

Here's what we built with Baseloop + Claude Code 👇

1. Every Monday, Baseloop pulls all follow-up tasks from HubSpot

2. It verifies the contact is still at the company
One in five aren't. Without this step, you're crafting personalized emails into the void.

3. It reads the full email thread and call notes. Then writes the follow-up.
Not "just checking in." A draft that references what was actually discussed last time.

4. The draft lands in lemlist
→ Vincent reviews and sends the first message
→ Lemlist takes over. Email, LinkedIn, phone.

40+ tasks. Zero research. Half a day back, every week.

HubSpot knows when to reach out. Baseloop reads the history and writes the email. lemlist gets it delivered.

How are you handling follow-up research right now?

Aurelien Aubert

Investor

ADHD + AI is an unfair advantage.

AI is the first technology that rewards ADHD traits.

Old model: focus on one task, finish it, move to the next.
Now: launch a fix agent, launch another for the docs, check the first output, correct it, rerun it.

You're not executing anymore.
You're orchestrating.

And if your brain naturally works that way, you're already ahead.

Aurelien Aubert 🧱

AiSDR just patented "generating emails with an LLM and responding to replies."

Good for them.

In 2023, that was genuinely exciting.
In 2026, every GTM tool does this out of the box.

If a patent is your competitive strategy, the product conversation is already lost.

Build a better product.
That beats any patent.

No hard feelings. Everyone plays their cards.

We just shipped the Cargo CLI.

2 weeks ago, we released our TypeScript SDK (@cargo-ai/api).
Today we're going one step further: All Cargo, from your terminal.

npm install -g @cargo-ai/cli

One command. Full access to your GTM infrastructure.

- Create a Tool. 
- Trigger a run. 
- Send a message to an agent. 
- Query your data models.

Why this matters:
→ Built for Claude Code, Cursor, and agentic workflows
→ Your AI coding assistant can now operate your entire GTM stack
→ No browser needed — script it, automate it, ship it

This is just the beginning. Exciting things coming.

https://lnkd.in/eGBVGsR7

I learned to code at 11.
I'm living the period I always dreamed of.

I can build faster than ever.
I feel like a kid again.

Went through the whole traditional path.
Engineering school, then dropped out.

Spent years building random apps.
Most of them ended up in the trash.
Did some freelancing to pay the bills.

Joined Spendesk early. 
I was an engineer back then.

Met Maxence on the growth team.
The GTM tools sucked.

Couldn't do what we needed.
So we rebuilt everything in-house.

That's when it clicked: we were power users who got frustrated enough to build our own solution.

That's usually the strongest signal there's a real product to make.

Went solo as a freelancer for a bit.
Tested things out.
Then Max and I decided to build it as a SaaS.

When you've already rebuilt the tools once because nothing else worked, you know exactly what's missing.

We just launched our npm package.

Official TypeScript client for the Cargo API: manage data models, run workflows, and build AI agents from Node.js. Use it in Cursor or Claude Code to trigger agents/tools, analyse data, or pull workflow and model analytics. One install, full types, less boilerplate.

https://lnkd.in/eyUWvNvX

The new test isn’t “does it run?”
It’s “does it survive production?”

You can vibe-code a GTM product in a weekend.

You can’t vibe-code:
- Connectors that don’t silently break
- Data that doesn’t duplicate or drift
- Retries that don’t corrupt state
- Permissions that don’t leak customer data

AI made code cheap
It also multiplied the surface area of failure.

Infrastructure becomes the product.

Architects design. Bricklayers execute.

In software, writing code is bricklaying.
AI handles it.
Designing systems is architecture.
AI doesn't.

Same thing is happening in GTM.
Shipping workflows, sequences, automations. Bricklaying.
Designing the system underneath. Architecture.

One is getting cheaper every day.
The other isn't going anywhere.

I talked with top SWEs in the Bay Area last month.

80% of their code is AI-generated.

They prompt, review, ship. That's the whole job now.

Then I talked with GTM engineers.

Still building Clay tables.

You have nothing of an engineer if that's your stack.

Vibe GTM is not a thing.

Stop trying to be a Clay expert.

Learn Claude Code. Write real code. Build real systems.

The ones who don't will be replaced by the ones who do.

AI increased the workload.

Nobody talks about that honestly.

14 automations. 8 integrations. 

Fridays spent debugging why the sequence broke.

Scale glue and you get more glue.

Engineers use AI to eliminate toil.

GTM teams use AI to stack more work on top of fragile systems.

Same hours. Different foundations.

Build vs buy? 

Building won.
But not the way people think.

AI made code cheap.
It didn’t make systems reliable.

So yes, teams will build their own GTM logic:
scoring, routing, enrichment, agents

Because that’s where the edge is.

But they won’t build:
- connectors that don’t break
- sync that doesn’t duplicate
- pipelines that retry safely
- schemas that don’t explode

That’s infrastructure.

GTM is going the way engineering already did:
- You own the code that differentiates you.
- You run it on infrastructure you don’t have to babysit.

If your GTM strategy is stacking SaaS tools, you don’t have leverage.
You have dependencies.

The next generation of GTM teams won’t have a better stack.
They’ll own their logic.

And run it on infrastructure built for GTM.

We’re all racing to build the next best thing in AI. 

But most of the world has never touched AI. 

Plumbers, dentists, local shops,… they’re running 2026 like it’s 2010. 

The tools exist. Nobody showed them. 

Master the gap. Sell it. Dominate. 

Big wins go to the ones who zoom out, not the first arrivals.

IBM dropped 10% the day AI proved it could refactor their code. 

If your product is just logic in a UI, and AI can rewrite it in a weekend, you don’t have a moat, you have a countdown. 

Claude refactoring COBOL isn’t hype. 

Forty years of “too risky to touch” just became accessible. 

Obscure stacks and legacy languages aren’t defensibility anymore. 

The companies holding value don’t own code. 

They own distribution, embedded workflows, and real operational data. 

The moat is the system around it.

Engineers built the thing replacing them. 

Agents run in parallel, ship faster than teams ever could. 

Integrations. Bugs. Data. Verification. 

But Taste decides what survives. 

No shitty products. No product without vision. 

AI gives speed. 

You give direction. 

That’s the combo that wins.

AI doesn't make all engineers better.
It makes the best ones far more productive. 

And the rest, replaceable.

10x engineer + AI = 100x engineer.
1x engineer + AI = 0.1x engineer.

The 100x runs agents in parallel while working on something else.
A full team without the team.

The 0.1x used to fix bugs, build integrations, ship simple features.
That's AI's job now.

The gap between engineers isn't closing.
It's compounding.

Clay's new pricing is a great deal. For Clay.

The problem is the limits.

40,000 data actions sounds big. It's not.
Here's what that actually means in the real world:

Enrich 10k leads with a 4-step waterfall. 40k actions gone.
Run a Clay agent on 5k records. You're already burning thousands more.
Sync data back to Salesforce, HubSpot, or your warehouse. More actions.

One decent GTM workflow and your entire monthly budget is gone.

Now the second illusion: "cheaper integrations."
Only a handful of actions actually got cheaper. The heavy usage stayed the same:
waterfall enrichment, people data enrichment, Clay agents, most serious GTM workflows.

So the headline pricing looks great.
But the moment you run real pipelines, the math explodes.

To be fair: splitting enrichment and orchestration is the right move. We did the same thing a year ago.
The problem is the implementation.

The people getting hit hardest are the exact ones who made Clay successful.
Agencies running enrichment at scale. Power users building complex tables.
Teams running multi-step pipelines. Companies pushing serious data volume.

The builders.

This pricing is optimized for enterprise contracts now. Which is fine.
But when a product starts optimizing for procurement, power users start doing the math.

If you're running high-volume enrichment, complex workflows, or a custom GTM stack, you'll feel this immediately.

To all my frenchies 🇫🇷

https://lnkd.in/eht5nBsa

Co-Founder @Youno | Faites de votre CRM un système de revenu prédictible | Clay & Cargo Partner 🧱

Votre CRM est peut-être votre pire ennemi commercial. 
Jeudi 5 mars avec Hugo, on vous montre comment en faire votre meilleur atout grâce à Cargo.

La plupart des CRM qu'on voit chez nos clients sont dans le même état.
Plein de data. Mais inutilisable.

Doublons, infos périmées ou manquantes, 
Sales qui perdent 15-20 min par fiche à reconstituer manuellement ce que l'outil devrait leur donner en 1 clic,
Head of sales qui répartit les comptes à l'aveugle.

Résultat : plus personne ne fait confiance à l'outil. 

On va vous montrer comment on a transformé le CRM d'un client en une source de données fiables et dynamiques avec Cargo 🧱.

Au programme :

→ Cargo vs Clay : pourquoi on a choisi Cargo pour ce projet et dans quels cas c'est le bon choix

→ Enrichissement + dédoublonnage : les workflows exacts qu'on a déployés pour nettoyer et faire vivre la base

→ Cartographie du TAM : segmentation en 3 tiers + 5 signaux d'achat activés (champion tracking, nouveaux recrutements, croissance…) 

→ Sales-ready : boutons HubSpot en 1 clic + agent IA sur Slack pour zéro friction au quotidien

À la fin de ce webinar, vous repartez avec une vision claire de ce qu'un CRM bien architecturé peut faire pour votre équipe commerciale.

📅 Jeudi 5 mars de 11h30 à 12h30
Lien d'inscription en commentaire 👇🏼

Julián MEDINA, Guillaume Josset and 6 other connections attending

Maintenir une data CRM propre et activable avec Cargo

taplio

The Battle

Robert Goncear

Aurelien Aubert

Followers

Posts

Engagements

Likes

Performance ScoreRobert Goncear's LinkedIn Performance

Performance ScoreAurelien Aubert's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Robert Goncear

Aurelien Aubert

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);}Robert Goncear's LinkedIn Performance

Performance ScoreAurelien Aubert's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreRobert Goncear's LinkedIn Performance