A self-building CRM is no longer a future idea - it's coming.

We’re moving from “filling CRMs” to “CRMs that build themselves.”

Just sat down with some founders to brainstorm arround this and we had this realization.

Everything needed for this is ready:

1. Voice-first input
Tools like Wispr Flow let you capture notes, updates, and context in real time - no more manual data entry.

2. AI-native GTM engines
Deepline is pushing toward an “AI-CLI for go-to-market” - enriching, qualifying, and triggering actions automatically.

3. Self updating context engines
Like updating your whole GTM-brain: offers, products, services, value props... based on what you say, what happens and your updates.

4. Self-evolving interfaces
Lovable-style UIs that could adapt based on usage, meaning your CRM doesn’t just store data - it restructures itself around how you work and your custom data needs.

5. Relationship intelligence layers
Platforms like The Swarm map actual human connections, not just static records - turning your CRM into a living network

6. API-first sequencing
Email + outbound tools are becoming composable - triggered via APIs, workflows, or MCP instead of rigid campaigns

7. Unified comms integrations
Unipile abstracts LinkedIn, email, and messaging into a single layer - making outreach programmable

8. Programmatic AD campaign platforms
Creating ad campaign proposals programmatically.

What this means:

The CRM will no longer be a system of record.
It’s becoming a system of action.
It should become an actual self-updating system.

Imagine your CRM like calling a collegue, and having it update your systems, signals and campaigns based on your business logic.

So, what's stopping this from becoming a reality?

Why not have a CRM that evolved as your business evolved?

The Future of Self-Building CRMs

"Clay changed its pricing. The era where you could save with technical workarounds (API connections, app integrations, etc) is over.  

Elaborated tables now correlate with increasing costs in Clay (whether you close those deals or not).

Let me explain what this means....

Now, similar to automation tools, actions you set up in the platform cost "action credits." That means you can run out of credits from data enrichment or the actions in your tables.

Actions are all value-add integrations that are not data enrichment from third-party providers: Slack messages, Notion integration, HubSpot lookups.

For many of the data enrichments, they've reduced the credit costs (which is good news).

Oh, and they removed the Explorer plan. Now the minimal entry point for API/HTTPS integrability is the $495/month plan.

So the stack cost just jumped for many teams using Clay for:
• enrichment
• signals
• AI personalization
• workflow automation

This happens at a moment when the elephant in the room—Claude Code—is gaining popularity among technical GTM Engineers.

They're right about one thing: Clay has evolved as a product from just data providing to offering tooling for end-to-end GTM workflows. Pricing is changing, and their logic is that value has also changed significantly; while having users over-leverage the platform with low-success for Clay.

That has changed the game, and now they had to make moves to adjust pricing. My respect on how they've been approaching it 👏 

I've gotten a lot of value out of it over the years and won't stop using it anytime soon - rather glad I'm staying on the legacy plans for a bit longer.

Yet for newcomers - this can be something to think about.

Last week I advised a leadgen agency moving their software from another tool into Clay... hoping Clay would be way more cost efficient. 

Imagine the surprise these folks will get when the so called "best product for GTM" now will keep growing costs as they scale success.

As many of you know, this ecosystem is going through a lot of disruption right now - and we will very soon see new players taking advantage of this opportunity with different, intuitive and AI-first offers.

For those AI-native GTM Engineers, the path forward is no longer clear with Clay - it's a signal to start exploring and upskilling yourself in new technologies.

For those growth and marketing teams who value speed and convenience:
→ get familiar with the changes 
→ lock in on those legacy-plans while still available
→ explore how to actually use Clay's new features like Sequencer, Intent and Agents. 

The real question is: will your current setup of Clay justify the cost increase? That's for you to decide

Clay Pricing Changes Explained

Took over a messy problem: fixing years of contract renewals relying on manual follow-ups. Two weeks later, we have a new ai-augmented process with a working system.

Here's how it works.

❕The Problem:

→ Renewals were everywhere and nowhere.

→ RevOps was supposed to check which contracts were renewing and chase the responsible person.

→ No notifications, no clear ownership assigned.

→ No process or automations built.

→ And worse: some contracts even quietly expired.

💡 The Solution

A process that assigns accountability and ensures no renewal is missed.

Actionable notifications. One tracking sheet. One automation system.

How it works:

Step 1: Contract tool sends a reminder

Months before a contract expires, the system gets alerted automatically. Instead of paying an expensive API - we capture the native notifications from the email to start the automation process.

Step 2: A case is logged

Every reminder becomes one tracked case. Company name. Contract date. Status. Owner assigned.

Step 3: The right owner is identified

AI enriches the company so we can find a match to a company in our CRM. Then AI assigns ownership on a renewal case to a commercial leader, based on predefined rules. If it matches - automatic assignment. If something's unclear - flagged for manual review.

Step 4: Owner gets a Slack message

A clear request with all information they need. "This renewal is yours. Confirm or reject ownership." And follow-up reminders until it's done or it reaches a deadline. The deadline... a few weeks before the contract expires.

Step 5: They take action

Talk to the customer. They fill out the final decision: renew or terminate. Or they can inform Legal of their decision - either way, the case will close.

Step 6: Manage the escalations

If the case owner doesn't respond? They rejected ownership? Escalation happens automatically. The case is alerted. It can get reassigned. And it's all shared in a dedicated Slack channel.

Every message is actionable, with Human-In-The-Loop option that allows the process to move forward:
→ Reassign case
→ Add a note
→ What's the resolution?

What changed with this approach?

Before: Hours finding owners. Renewals lost between systems. One person held all the knowledge.

Now: Automatic routing. Everything tracked in one place. The process knows.

The Results

✓ Zero missed renewals
✓ Zero manual chasing
✓ Clear accountability on day one
✓ Full legal visibility
✓ Automatic escalation
✓ Structured follow-up

The system isn't complicated. It's just clear.

It chases down everyone until a case is closed.

That's where RevOps makes a dent. Not in being optimistic about people, but in building systems that ensure things get done.


Oh and the engineering of it? 

Not trivial - I used to charge clients thousands for these builds - but it gets easier as you know what you need and have the right tools to build.  In my case, I did the magic with relay.app/?via=ivan

I've been using it since '23 - don't wait another year to try it 👀 (Link below)


How I Solved Contract Renewals For Revops & Legal

I wasted months trying to fix a broken content workflow.
This weekend I rebuilt it from scratch in 2026 standards.

I struggled with content production for years.
The writing wasn't the hard part, the system behind it was.

My old content system was broken.

It used to live in an Airtable and powered by clunky automations. The interface never sticked to me, it was disconnected from my other tools... at some point I gave up.

This weekend I took a new path.
I upgraded everything to 2026 standards.

I stopped making tools fit my needs.
I started building my own custom tools.

This is the power of vibe coding your own tools.

I started with my database structure.

1. I moved my past setups to a tool called @Zite.
Zite is an incredible tool where vibe coding meets "no-code" intuitiveness.

2. I used the databases MCP to connect it to Claude.
This allowed me to develop the database that now powers my content engine.

3. With Claude, I built the source prompt for my application (V1) in Zite.

A Kanban board became my content pipeline visualization.

Every step on the board triggers something.
The pipeline moves from idea to generation.
It goes from drafts to final editions.
Ready to schedule and post.

Everything is tailored to my content pillars.
It fits my funnel and post styles.

4. Data flows through different stages into a tool called Relay.
I set up Relay to use AI agents that generate and edit content like me.

AI augments my process, it doesn't replace it.

5. Then I built the features I wanted on top of my engine.

- I can pick specific post templates from a gallery of top-performing formats.
- I can integrate transcripts from my meetings and voice recordings to create content ideas.
- I even connected it to Typefully to send content drafts directly into LinkedIn and Twitter.

Vibe coding built the app in hours.
The framework behind it took years to assemble.

Prompting made this possible faster than ever.
I put version one together in hours.

Then I added features at insane speeds.

I haven't felt this excited since 2019, when I discovered no-code tools.

In fact, this post was written from the content engine.
It's live for me now.

And I'm done gatekeeping my creations.

The future of software is building your own.
That's why I'm sprinting on to something new.

I'm creating ready-to-deploy GTM tools. 

Taking up a personal challenge: Building 8 GTM tools in 10 weeks.

Tools that will include full backend, tool builder, and automations - which I'll be able to install to any client in a single call, with no Saas-lock-in, and built with intuitive, easy-to-edit softwares.

This was system #1. Want your own content engine? 

Let's build your own (I'm taking first pilot customers).

Building a B2B Content Copilot

Clay was failing me - so I build my own deep-enrichment tool.

Back in December, when Gemini 3.0 Pro had just launched, I ran into a problem while doing company segmentation at scale

I was processing thousands of companies in the energy sector and needed deep website research to structure data about their projects.

Naturally, I tried running it through Clay.

But I hit a few limits:

→ Deep research agents were expensive at scale
→ Gemini 3.0 wasn’t available yet through integrations
→ Long-running research jobs would struggle with rate limits
→ And worst of all: they didn't have a way to "delay/wait for minutes" before getting data from an API, so I'd be impossible to run deep-research agents that take 1-3 minutes.

None of this is a criticism of Clay. I use it every day, and give them feedback often.

By now, Gemini 3.0 is available in Claygent and the Delay feature was released!

But as a GTME, when a tool hits a constraint, you can't just stop your GTM. I asked myself:

What part of this workflow can I rebuild myself?

So I ran a quick experiment.

Using Google AI Studio, I built a small enrichment layer that could:

→ Import company data from CSV
→ Inject variables like company name and domain
→ Run structured research with Gemini 3.0 Pro
→ Return clean JSON outputs
→ Wait as long as the deep research needed
→ Export results back to CSV or push them into Clay via webhook

In other words: a custom research agent for one specific job.

Not a replacement for Clay.

Just an internal tool that handled the part where my workflow needed something different.

The results from this experiment we're quite impressive as well - Gemini 3.0 Pro with "Grounding Search" turned out to have VERY accurate results, with no hallucinations or created URLs, yet with a very short depth in what it would find (compared to Exa or Claygent, which will find more results over the surface).

The interesting part is how easy this has become.

With AI-assisted development, building small internal tools like this now takes hours instead of weeks.

So the mindset shift is simple:

If a tool doesn’t support your workflow yet, don’t just wait for the feature and block your GTM initiative.

As a system thinker, first try to understand why the constraint exists.

Ask about it.

And if no solution is available, see if you can extend the system yourself.

Sometimes the fastest way to ship isn’t switching tools.

If your tools fail you, try building around them.

Curious: what internal GTM tooling have you built yourself?

Would love to hear examples 👇

Building a Deep-Enrichment Tool

Just spent a week at MWC, 4YFN and Talent Arena in Barcelona. As a new mover to Spain, I wasn’t sure what to expect from the tech ecosystem here. After this week, I confirmed what I suspected: 

Barcelona’s tech scene is on another level!
And this week went way further than Spain, it was truly global.

MWC had 105,000 visitors from 207 countries, 2,900 exhibitors and 1,200+ speakers. Talent Arena had 25,000 attendees. There were ecosystem hubs full of startups, a massive "XPRO" zone for advanced devs, AI conferences everywhere, YC startups and top VCs all around.

What surprised me most? How organized everything was around ecosystems, industries, and countries.

Startups weren’t scattered. They were grouped by region and niche, making it easy to discover companies you'd never find otherwise.

The agenda, logistics and app experience were also smooth - no complications.

Builders came from everywhere!

I spent time with Juan Velasquez (AI engineer, fellow maker and former Overdrive collaborator) bouncing between AI engineering sessions and networking events.

Some standout moments:

→ Benedict Evans on “AI Eats the World” (perfect title for the whole vibe)

→ Dealroom's networking event, connecting with one of the top data providers for VCs

→ Marietje Schaake talking about Europe’s path to a digital single market and the geopolitical shifts shaping it

→ Carlos Milan from AWS breaking down MCP security risks (made me rethink how casually I install MCP servers)

→ Steve Aoki explaining how AI changed the game in music production

And I had one serendipitous find.

Met the team from Biyiud – Sustainability as Wealth, a climate-tech startup from Barcelona. Their model is almost identical to Ecobyte, the sustainability incentive platform I launched in Colombia in 2021 (my first startup).

Same thesis, same targets (consumers, companies, cities) — but in a more mature market.

Wild coincidence, and an even bigger surprise: the founder is Colombian. 👏

Some other startups that caught my eye from a GTM lens:

→ Glink (YC) – Clay-for-healthcare. Interesting for healthcare sales
→ GeekSource (Netherlands) – Lemlist-like tool for automating candidate sourcing
→ vtouch (Korea) – AI voice ring. Perfect for conferences and sales conversations
→ Spendbase (Ukraine) – Optimizes Cloud & SaaS spend + simplifies LLM key management. They even have a team in Medellín

Sharing more in the comments ↓

One thing I didn’t love: access restrictions.

Too many passes: VIP, investor, industry-specific, partner-only. Some events required email approval on top. Felt overly gatekept compared to other tech conferences where you can roam freely.

But overall?

This week helped me *see* how alive Spain and Cataluña’s ecosystem really is.

The infrastructure, the community, the momentum - it’s all there.

PS: I’m still new to the city, so if you’re in the Barcelona GTM & startups world… who should I meet?

MWC and Talent Arena first impressions

This is the most insane vibe-coding design workflow I have found. 

I started working on my wedding invitations. A simple personal project - an addition to my wedding site.

I wanted something visually stunning. Not basic. Not ai-sloppy with purple gradients.

Here's where I was stuck: design inspiration was scattered... 
- Pinterest boards. Dribbble, Instagram
- Figma design references
- Screenshots

But translating that into actual code? 
Translating into instructions for Lovable/Zite/Claude for this specific project?

That was the real problem. I found good options but nothing actionable to start this project.

And sure, I'm familiar with Figma - and because of that, I didn't want to go into that rabbit hole.

Then I found a different approach.

1. I picked my favorite reference I liked on Instagram and took a few screenshots of it

2. Then I uploaded them into this tool called Variant with a prompt describing what I wanted.

@GSekretaryuk discovered this tool the other day - thanks for the tip!

Variant did the 0-to-1 creative heavy lift, making a feed of design options.

It understands what you want and generates multiple design options automatically, with different styles. 

So instead of doom scrolling through design options, I could now scroll through ready-to-go designs.

Senior design work, instantly made, and with multiple "variants" to pick from (points for smart branding!).

I picked the version I liked most. The colors. The layout. The copy.

Then I copied the code and went to Bolt > pasted it to build this new section of my website.

The entire process from start to finish must've taken an hour at most. 

The wedding invitation's feature of my site is now live, and not to brag - but it looks amazing! We're already sending them out to guests.

As an app builder and a former product-person - what I just did felt insane.

I became the creative director with minimal friction.

I pointed. I picked. I refined.

The AI handled the execution.

An absolute shift. 

Anyone can turn their ideas into beautiful creations now, which go WAY past the "ai slop" and "purple gradients" that we see all over the internet since Lovable became a thing 🙄

If you're building software, you have to try this workflow out.

Insert your idea. Get different options. Choose what works.

Whether it's a wedding site or a new feature for your platform.

Going from design to building is now faster than you've ever thought possible.

And the people who understand these 0-to-1 workflows? 

They will be moving significantly faster than everyone else (something worth thinking about!)

Vibe-Coding Wedding Invitations Like A Pro

We had less than 20 days to plan a wedding in a Swiss castle... so we naturally proceeded to ship a wedding website in a few days 🤷🏽‍♂️

(This is NOT a “what getting married taught me about B2B sales” post lol - but it’s a crazy story that might get you into giving vibe coding a chance.)

You see... we initially thought we’d get married in a government office.

Then we discovered we could get married in a swiss castle.

Applied.
Got approved. Hell yes!
Then we immediately realized: “Oh… sh*t, we’re doing this. In less than 20 days. OMG!”

So, naturally, we went full scrappy startup mode 🤣

→ Planned the wedding with ChatGPT + a simple Notion setup

→ Built the site with Bolt (event pages, itinerary, guest quiz, registry, admin view)

→ Researched and found all our vendors!

But I had been thinking of making a wedding site for a while, so we got on it!

We wanted it on our own way: instead of a traditional registry, we'd let guests support our wedding & honeymoon experiences (we're a bit... nomadic)

We built the site with Bolt (event pages, itinerary, guest quiz, registry, admin view)

→ Launched a very rough V1… the next day!

Then reality hit:

“We can’t use Stripe or Lemonsqueezy, we’re not a business”
“How do we send emails?”
“Where do we even track gifts?”

So at 10pm, I did what I do best: I built automations 🕶️

I connected Bolt to Relay with webhooks so that:

→ When someone requested bank details, we sent them only after a human-in-the-loop approval

→ When someone chose a gift, we sent a personalized thank-you email, approved before sending

→ Everything was logged into Google Sheets so we could actually track what was happening

...It was go time 🥁 

We launched the site to our guests at 11pm, just 15 days before the wedding.

And to our surprise, it was working, and people liked it!

People started using the site. 

Gifts came in.

Bugs and content issues appeared of course, but that's fine, you can fix them quick. 

And fiancé became my QA! 

“This is confusing” → fix → ship → repeat.

And best of all, the site became the source of truth for our wedding.

By the big day, everyone knew where to go, what to do, and what was happening.

But honestly?
The site was just really fun to build together.

It’s the first thing we ever "shipped" as a couple.

We planned together. We built together.

And yes, we did get married in a Swiss castle, was very dreamy🏰! 

10/10 would recommend. 

PS: Vibe-code your wedding site - or any personal project. You’ll learn a ton and have way more fun than you expect.