Marcel Santilli

CEO/cofounder @GrowthXAI - AI powered growth 🇧🇷🇺🇸🚀 // Previously CMO/growth at @DeepgramAI @scale_ai @HashiCorp @ServiceTitan

[Marcel Santilli @santilli]

BIG NEWS: We just launched CheckThat[.]ai

Buyer Intent + Free AEO Platform + Shared Prompt Library

No cold start. Historical data included.

---

"Seriously!? Yet another AEO tool?"

I get it. I felt the same way.

But I'm convinced the AEO category is fundamentally broken.

Not the tools themselves. The assumption behind ALL of them:

That tracking = improving.

It DOES NOT.

Google Analytics doesn't grow your traffic.
An AEO subscription doesn't grow your AI visibility.

What actually happens:

Sign up → bad context → random prompts → zero data → $500-$3K/mo invoice.

For prompts you're not even confident in.

---

We took the complete OPPOSITE approach with CheckThat:

We did the heavy lifting before you show up:

→ Top B2B software categories mapped
→ 5,800+ brands researched (and growing)
→ Millions of AI responses tracked daily
→ Human-reviewed prompt library (shared + free)

Why?

Working with 100+ brands at @GrowthXAI taught us:

CONTEXT is everything.

Your context drives your content quality.
Your content quality drives your AI visibility.

You can't create great content without context.

You can't get context without understanding your buyers.

You can't understand your buyers without data.

---

To celebrate the launch, I'm doing
50 Free AEO Audit + Strategy sessions over the next month

Retweet this post & comment "checkthat" and I'll send you a DM to get scheduled.

[Marcel Santilli @santilli]

@michlimlim Loved it!

Stay close to customer. Turn hands on customer tasks into product. Post consistently on Linkedin. Turn buyer conversations into posts.

[Marcel Santilli @santilli]

@ConorBronsdon Pretty cool, I gotta check it out

[Marcel Santilli @santilli]

@Shyamal890 @JHTScherck @ahrefs @semrush @dataforseo I did a 90min workshop on this in case it's helpful https://t.co/1ZD18WG2Os

[Marcel Santilli @santilli]

@ConorBronsdon @cclark So true!! The worse is when it feels like people are trying to delegate their thinking instead of using AI to LEARN FASTER.

[Conor Bronsdon @ConorBronsdon]

+ check out the full @chain_ofthought conversation with @santilli on building an AI-native startup (and how Marcel did it!):
https://t.co/oF78kWIOZg

[Marcel Santilli @santilli]

@ConorBronsdon Thanks for having me on

[Marcel Santilli @santilli]

Today is my birthday, and I keep thinking about a school cafeteria table.

I was 11 when we moved from Brazil. I didn’t speak English. I was a stranger in what felt like a completely different universe.

School became a daily exercise in trying not to be laughed at. Trying to blend in. Trying to get through lunch.

I can still see the cafeteria. It seemed like everyone had a table. Everyone knew where to go.

I didn’t.

I think some part of me has been standing in that room ever since, looking for a place to belong.

For a long time, my answer was to work harder. Outsmart everyone who wouldn’t let me in. I watched people closely. I learned to spot patterns, think ahead, figure out how to get their attention.

Those skills have helped me build a life. They also came from a kid who was scared.

Then I learned graphic and web design. I could sit for hours making things, like I used to with Lego. I built a website and figured out how to get customers online.

The first time a real business filled out my form and paid me for work, I was blown away. I was a teenager. They had no idea.

I’d made something and someone found it. For once, I didn’t have to wait for someone to pick me.

I can see that thread through so much of my life now. Building things people might care about. Giving away value. Hoping the right people would find their way to me.

MY PEOPLE.

I wanted more than customers. I wanted a team I loved being around. I wanted to make a place where we could do work we believed in and where other people felt they belonged too.

My team and I are building that table at @GrowthXAI.

We help companies figure out what they stand for, make something valuable, and find the people they’re here to serve.

I think of a website as their TABLE: a place to show up as themselves and invite those people in.

And honestly, the fear hasn’t gone away. I still wonder if I’m faking it. If I’ll fail and everyone will see. I know where that voice came from. I’m trying to hear it without letting it run my life.

…

Today, I just want to express my gratitu to the people around me. Those that have joined my table in some way. At the team building this with me. At everyone who made room for me along the way.

Thank you. It means the world to me.

Much love ❤️

[Marcel Santilli @santilli]

@mgonto @guillaumecabane Epic!! Would love to see you if you're around next week.

[Marcel Santilli @santilli]

@jasonlk I found more success looping through with multiple test set. After the 5th loop adjusting prompts and criteria it got a lot better. Also you can run several classification and then jev-as-a-judge.

Also this is insanely powerful https://t.co/0mmIHCQzao

[Marcel Santilli @santilli]

@jasonfried A good way to know is your level of frustration working for every company. If manageable you're keep working for someone. I found myself so frustrated all the time. Turns out it was a me thing. And I worked for some amazing companies.

[Marcel Santilli @santilli]

Super balanced take. We've observed similar patterns. We have several millions AI responses tracked monthly for high quality evaluation stage questions with citations scraped. If you ever want to partner on something lmk.

One thing that we found makes a difference is to map the ecosystem, buying categories, adjacent buying categories, buyer criteria, and use cases. Then take these context modifiers and create valuable content around them. Especially in AI mode, controlling more of the citations help.

Another thing that brands forget is how much training data anchors these answers, especially for Claude and ChatGPT.

[Marcel Santilli @santilli]

that's exactly why the future for *most* knowledge work is more like how engineering works today. Shared context, shared memory, agentic workflows as code centralized from an approved library.

We use our own agentic framework (https://t.co/EuIyJZkVSh) w/ durable execution/evals/tracing. It's way more scalable

[Marcel Santilli @santilli]

@Altimor @thomaspower Looking forward to reading more about it. Curious if you all built the infra in a way that makes it easy to model switching in run or have evals at scale and fall backs.

We tried to build some of these primitives into https://t.co/iwhwbJxMZq (OSS)

[Marcel Santilli @santilli]

https://t.co/pAZRZXgvVS

[Marcel Santilli @santilli]

Survey is live. 27 questions on AI adoption, budget, content ops, and AI visibility. Not asking if you use AI -- asking how your strategy has actually changed. Link in comments.

[Marcel Santilli @santilli]

@nbaschez Been using Obsidian but def not ideal experience back and forth. Curious

[Marcel Santilli @santilli]

@dharmesh We built https://t.co/iwhwbJxMZq as a typescript agentic workflow framework, Claude Code native :)

[Marcel Santilli @santilli]

That's why we built and open sources https://t.co/EuIyJZko2J

As we were scaling service as software for https://t.co/s3WGJgaTnA we needed something that was like what Terraform was for infra but for knowledge work.

Agentic workflows as code
Evals. Tracing, cost tracking , human reviews
Durable execution (temporal abstraction)
Plug and play with coding agents

Feedback welcome

[Marcel Santilli @santilli]

Skills are great for things that are not meant to be shared, don't require more structure input / output, doesn't require long durable execution, where traces are not as important.

If you think of skills as knowledge work as written spec that needs to be "run" consistently I think you end up needing agentic workflows with things like scenarios, traces, evals, runtime etc.

https://t.co/dwZN7l7aqY

[Marcel Santilli @santilli]

@garrytan That's partly why we built and open sources https://t.co/iwhwbJxMZq

As we were building https://t.co/KMsabZhnuR (service as software) we felt like we needed:

Agentic workflows as code (Claude Code native)
Evals, tracing, llm as judge, cost tracking
Durable execution (run time)