Training on AI and automation for business professionals: workshops, online courses, live sessions, all the way down to custom company training programs

Co-Founder

Circula

Advisor

Circula.com

Head of Operations

On Deck

Fellow

solarisBank AG

Senior Business Development Manager

payleven Global Mobile Payments

IT Team Lead

Fundraising - Internship

Helping startup operators move faster with automation

I train business professionals on AI and automation: workshops, online courses, live sessions, all the way down to custom company training programs

Lindy 3.0 is absolutely wild. I just prompted - 𝘢𝘯𝘥 𝘵𝘦𝘴𝘵𝘦𝘥 - an agent that grabs a reply in Airtable and posts it on LinkedIn using Lindy Computer.

Yesterday's release is a breakthrough in two areas:
1️⃣  Agent builder further democratizes automation
2️⃣  Lindy Computer extends automation possibilities

Here's why:

𝐀𝐠𝐞𝐧𝐭 𝐛𝐮𝐢𝐥𝐝𝐞𝐫 𝐟𝐮𝐫𝐭𝐡𝐞𝐫 𝐝𝐞𝐦𝐨𝐜𝐫𝐚𝐭𝐢𝐳𝐞𝐬 𝐚𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐨𝐧. 

Lindy's agent builder is a step forward in enabling more people to build automations.

It asks clarifying questions to really nail your requirements, and will only start building once the plan is clear.

This is not the first agent builder, but somehow this one is the first - with Airtable Omni in a different category - that I can 𝘵𝘳𝘶𝘴𝘵 with creating something that may just work from the prompt I write.

The big advantage Lindy has: an LLM "sits on top of every module" and is guessing the settings. And the guesses have gotten pretty accurate lately.

For simple use cases, there's barely anything to touch. You can turn it on and it works. 

This means: many more people can now build their first AI agent. 



𝐋𝐢𝐧𝐝𝐲 𝐂𝐨𝐦𝐩𝐮𝐭𝐞𝐫 𝐞𝐱𝐭𝐞𝐧𝐝𝐬 𝐚𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐨𝐧 𝐩𝐨𝐬𝐬𝐢𝐛𝐢𝐥𝐢𝐭𝐢𝐞𝐬

This is the most exciting release of the year for me. 

Agentic process automation (APA) is the technological advance I'm most looking forward to. 

If there was an API for everything, we wouldn't need it at all. 

But there 𝘪𝘴𝘯'𝘵 an API for everything. 

And so there are millions of tasks that can't be automated the "standard" way.

The promise of AI computer use: automating tasks that can only be performed on the browser so far.

Here again I've got my hands on both Lindy Computer and ChatGPT Agent and the difference is day and night.

ChatGPT Agent loops for 10 minutes and comes back empty handed.

Lindy is more thoughtful and nails my assignments. 

It just works, and I can't wait to explore more advanced use cases with it. 


𝐒𝐚𝐦𝐞 𝐬𝐚𝐦𝐞. 𝐁𝐮𝐭 𝐭𝐡𝐢𝐬 𝐭𝐢𝐦𝐞, 𝐢𝐭 𝐰𝐨𝐫𝐤𝐬

This is not the first agent builder. 
This is not the first computer use agent. 

But so far these features missed the reliability mark. 

Lindy just reached it.

Kudos Flo Crivello and team 🫡

Alexandre Kantjas

There is 𝘰𝘯𝘦 𝘵𝘰𝘰𝘭 I believe absolutely 𝘦𝘷𝘦𝘳𝘺𝘰𝘯𝘦 should be using and that will completely change how (fast) you work:

Not ChatGPT or Claude.
Not n8n, Make, Zapier or Lindy.
Heck no, not even Lovable or Cursor!

It's much easier to use than any of these...

It takes 5 minutes to setup...

It does ONE thing...

...but it does SO WELL.

I can't see myself going back to not using Wispr Flow.

Don't type: just speak. I'm done typing with my keyboard 90% of the time. They claim to be "4x faster than typing" - and I buy it.

This is the (first) voice AI moment I've been waiting for so long. The quality + convenience of the voice dictation from this tool are just insane.

If you aren't using it yet - check them out.

P.S: I dictated this post, just made a few cosmetic changes in the end.
P.P.S: I am not affiliated with Wispr Flow, just a happy user.

n8n is amazing, but let's be honest: the learning curve is steep. Especially if you want to build 𝘢𝘨𝘦𝘯𝘵𝘴 with it. But there's a 𝘴𝘩𝘰𝘳𝘵𝘤𝘶𝘵...

That shortcut: Model Context Protocol (MCP).

MCP allows you to generate n8n AI agents from a prompt sent to Claude.

---
WORKFLOW BRO DISCLAIMER [START]
- This is not a magic solution 
- You will still need to create your connections
- You will still need to make some adjustments
- There's a bit of configuration required before you can start prompting agents 

But it's a 𝘮𝘢𝘴𝘴𝘪𝘷𝘦 𝘴𝘵𝘦𝘱 𝘧𝘰𝘳𝘸𝘢𝘳𝘥 compared to starting from scratch. And it will save you hours of building.

WORKFLOW BRO DISCLAIMER [END]
---

So that makes this topic perfect for... a 𝘭𝘪𝘷𝘦 𝘸𝘰𝘳𝘬𝘴𝘩𝘰𝘱: 

𝐁𝐮𝐢𝐥𝐝 𝐧𝟖𝐧 𝐀𝐈 𝐀𝐠𝐞𝐧𝐭𝐬 𝐰𝐢𝐭𝐡 𝐂𝐥𝐚𝐮𝐝𝐞 & 𝐌𝐂𝐏
👤 𝐇𝐨𝐬𝐭: Jan Meinecke
🕒 𝐓𝐢𝐦𝐞: Thursday, August 14th, 5:00pm CET / 11:00am EDT
📍 𝐋𝐨𝐜𝐚𝐭𝐢𝐨𝐧: Online

In this session, Jan will demonstrate how you can build this setup and how to overcome the remaining technical challenges you will likely face.

Interested? Drop me a comment and I'll share with you the link to register to the event.

9k subscribers for 9x on YouTube!

This is only the beginning…

No alternative text description for this image

AI agents are great - until you realize how 𝘶𝘯𝘱𝘳𝘦𝘥𝘪𝘤𝘵𝘢𝘣𝘭𝘦 they are. It’s not a bug. It’s their core feature:

We’ve been trained to think adaptability is always good. But when you’re automating ops, 𝘶𝘯𝘱𝘳𝘦𝘥𝘪𝘤𝘵𝘢𝘣𝘪𝘭𝘪𝘵𝘺 𝘪𝘴 𝘢 𝘭𝘪𝘢𝘣𝘪𝘭𝘪𝘵𝘺.

Most operational processes don’t need creative thinking or improvisation.
They need to work. Every. Single. Time.

That’s where AI workflows shine:
→ Fixed steps
→ Known outcomes
→ Boring by design

AI agents on the other hand shine in their adaptability. They can "think for themselves". Sometimes it's brilliant, most times it's chaotic.

But because they're not 100% error-proof, I would not use them in most cases.

Bottom line: what does your use case needs the most?
- If you want predictability, use workflows.
- If you want adaptability, use agents.

Just don’t expect you’re getting both with agents.

P.S: I choose predictability 90% of the time.

If the models are getting 2-3× better each year, why hasn’t 𝘰𝘶𝘳 𝘸𝘢𝘺 𝘰𝘧 𝘸𝘰𝘳𝘬𝘪𝘯𝘨 improved at the same rate?

This week, I felt a bigger gap than ever between the 𝘩𝘺𝘱𝘦 around new model releases and the 𝘳𝘦𝘢𝘭𝘪𝘵𝘺.

FOMO is strong.
The demos are weak.

We need less:
→ Benchmark charts nobody outside research cares about
→ Isolated model features that don’t integrate anywhere
→ Coding showcases (useful for a fraction of people)
→ Demos only computer scientists marvel at
→ “Our new model is now {insert stat}”

And we need more:
→ Case studies from normal teams, not just elite AI labs
→ Models integrated into the software we already use
→ Clear, replicable use cases for business workflows
→ People showing their craft with AI in real contexts
→ Demos of full systems solving real problems

We don’t care about more intelligence.
We don’t need more speed.
We need more 𝘶𝘴𝘦.

Let’s bring the question of 𝘩𝘰𝘸 𝘵𝘰 𝘰𝘱𝘦𝘳𝘢𝘵𝘪𝘰𝘯𝘢𝘭𝘪𝘻𝘦 𝘈𝘐 where it belongs: front and center.

I am both amazed and puzzled by the GPT-5 demos showcased since yesterday.

Amazed: looks like we'll all soon be able to vibe code or generate single usage interfaces directly from LLMs.

Puzzled: I have no need for that, and LLMs still can't do my work.

I really feel a disconnect between the demos and my (simple) needs. 

Am I the only one?

AI is removing the biggest blocker for non-technical builders: 𝘭𝘦𝘢𝘳𝘯𝘪𝘯𝘨 𝘴𝘺𝘯𝘵𝘢𝘹.

Before, building automations felt like learning kung fu the hard way. Endless hours studying obscure formulas, JSON paths, bracket nesting.

You needed to spend considerable time on the (hopefully clear and up-to-date) documentation.

One mistake and it all broke.

This discouraged many wannabe builders.

Now you only need to:
a) 𝘬𝘯𝘰𝘸 𝘸𝘩𝘦𝘳𝘦 𝘵𝘩𝘦 𝘪𝘯𝘧𝘰 𝘪𝘴, and
b) 𝘧𝘦𝘦𝘥 𝘪𝘵 𝘵𝘰 𝘈𝘐.

AI can process new info like Neo in 𝘛𝘩𝘦 𝘔𝘢𝘵𝘳𝘪𝘹 - thanks to in-context learning.

And for you, this means: no more syntax fear.

You just say what you want:

“𝘊𝘢𝘭𝘤𝘶𝘭𝘢𝘵𝘦 𝘥𝘢𝘺𝘴 𝘣𝘦𝘵𝘸𝘦𝘦𝘯 𝘰𝘳𝘥𝘦𝘳 𝘢𝘯𝘥 𝘥𝘦𝘭𝘪𝘷𝘦𝘳𝘺” → Airtable writes DATETIME_DIFF({Delivery Date}, {Order Date}, 'days')

“𝘍𝘪𝘹 𝘵𝘩𝘪𝘴 𝘍𝘓𝘖𝘞 𝘧𝘰𝘳𝘮𝘶𝘭𝘢 𝘪𝘵’𝘴 𝘳𝘦𝘵𝘶𝘳𝘯𝘪𝘯𝘨 𝘦𝘳𝘳𝘰𝘳𝘴” → AI finds the misplaced parenthesis and rewrites:
IF({Revenue} > 0, {Revenue} / {Leads}, 0)

"𝘈𝘭𝘴𝘰 𝘢𝘥𝘥 𝘢 𝘤𝘰𝘯𝘥𝘪𝘵𝘪𝘰𝘯 𝘵𝘰 𝘦𝘹𝘤𝘭𝘶𝘥𝘦 𝘧𝘳𝘦𝘦 𝘵𝘳𝘪𝘢𝘭𝘴” → AI adds:
IF(AND({Revenue} > 0, {Plan} != 'Free Trial'), {Revenue} / {Leads}, 0)

“𝘏𝘦𝘳𝘦’𝘴 𝘵𝘩𝘦 𝘈𝘗𝘐 𝘥𝘰𝘤𝘴, 𝘣𝘶𝘪𝘭𝘥 𝘮𝘦 𝘢 𝘗𝘖𝘚𝘛 𝘳𝘦𝘲𝘶𝘦𝘴𝘵 𝘵𝘰 𝘤𝘳𝘦𝘢𝘵𝘦 𝘢 𝘯𝘦𝘸 𝘤𝘰𝘯𝘵𝘢𝘤𝘵” →
AI reads the JSON schema, outputs a fully formed curl request or a fetch() function with headers, auth, and payload ready to run.

“𝘖𝘯𝘭𝘺 𝘪𝘯𝘤𝘭𝘶𝘥𝘦 𝘱𝘢𝘪𝘥 𝘪𝘯𝘷𝘰𝘪𝘤𝘦𝘴 𝘧𝘳𝘰𝘮 𝘭𝘢𝘴𝘵 𝘮𝘰𝘯𝘵𝘩” → Make locates bundle[1].line_items[].status, builds a filter using formatDate and lastMonth() functions, no digging required.

AI handles the structure. You focus on the problem.

The bottleneck isn’t “How do I write this?”

It’s become: “What should I build?”

If this week‘s news about Google indexing ChatGPT convos taught me one important thing it‘s this:

You can’t trust platforms to protect your data.

Both OpenAI and Google fumbled it - badly.

Here’s what happened:
- OpenAI let you create public share links for ChatGPT chats.
- Those pages were accessible without login.
- They didn’t add proper noindex tags or headers.
- Google’s crawler found them - and indexed them.

Result: Private or semi-private AI chats started showing up in Google search.

Let’s be clear:
- OpenAI’s mistake: Naive defaults. Weak metadata. No serious safeguards.
- Google’s mistake: Crawled and surfaced content it shouldn’t have.

Both operated on optimism over caution, and users paid the price.

The broader lesson:

If something is viewable without logging in, assume it’s public.

If it’s public, assume it’s indexed.

If it’s indexed, assume it’s permanent.

Don’t rely on robots.txt.
Don’t assume “unlisted” = private.
Don’t trust the platform. Ever.

This isn’t about AI - it’s about basic digital hygiene in 2025.

And clearly, even the giants forgot to wash their hands.

AI regulation - especially the 𝘌𝘜 𝘈𝘐 𝘈𝘤𝘵 in the European Union - is coming in 𝘧𝘶𝘭𝘭 𝘧𝘰𝘳𝘤𝘦. What are your obligations if you're using AI?

Just finished recording a lesson on my favourite topic: 𝘈𝘐 𝘳𝘦𝘨𝘶𝘭𝘢𝘵𝘪𝘰𝘯𝘴.

Most businesses using AI tools are flying blind on compliance. 

And that's a ticking time bomb.

The EU AI Act isn't just for tech giants. If you're using AI for 𝘩𝘪𝘳𝘪𝘯𝘨, 𝘤𝘶𝘴𝘵𝘰𝘮𝘦𝘳 𝘴𝘦𝘳𝘷𝘪𝘤𝘦, or 𝘰𝘱𝘦𝘳𝘢𝘵𝘪𝘰𝘯𝘢𝘭 𝘢𝘶𝘵𝘰𝘮𝘢𝘵𝘪𝘰𝘯 (hello, that's basically everyone), you need to know where you stand.

The breakdown is actually pretty straightforward:
🔴 𝐏𝐫𝐨𝐡𝐢𝐛𝐢𝐭𝐞𝐝 = Just don't. Think social scoring or manipulative AI
🟠 𝐇𝐢𝐠𝐡-𝐫𝐢𝐬𝐤 = Heavy compliance (HR tools, credit scoring)
🟡 𝐋𝐢𝐦𝐢𝐭𝐞𝐝 𝐫𝐢𝐬𝐤 = Transparency obligations (chatbots, content generation)
🟢 𝐌𝐢𝐧𝐢𝐦𝐚𝐥 𝐫𝐢𝐬𝐤 = You're good, just follow best practices

Most business automation falls into Limited or Minimal risk - but you STILL need documentation showing you've assessed this. Ignorance isn't a defense when auditors come knocking.

So start documenting your AI use cases NOW. 

P.S: This lesson drops next week in our 𝘈𝘐 𝘖𝘱𝘦𝘳𝘢𝘵𝘰𝘳 𝘤𝘰𝘶𝘳𝘴𝘦.  I tried to make compliance actually digestible (yes, really).

Recording our lesson on 𝘈𝘐 𝘌𝘵𝘩𝘪𝘤𝘴 tomorrow. Here's a teaser:

With AI, we're not just moving data anymore.

AI is deciding who gets the loan. Who gets hired. Which patients get seen first. What content shapes your customers' worldview.

The invisible hand of algorithms is already writing the script of human lives - and most businesses are flying blind.

𝐄𝐯𝐞𝐫𝐲 𝐚𝐥𝐠𝐨𝐫𝐢𝐭𝐡𝐦 𝐭𝐨𝐮𝐜𝐡𝐞𝐬 𝐚 𝐡𝐮𝐦𝐚𝐧 𝐥𝐢𝐟𝐞.

In this lesson, we'll be breaking down the 8 ethical landmines that can either sink your AI projects or become your biggest competitive advantage:
→ Bias that builds unfair systems 
→ Copyright chaos in training data
→ Privacy laws that can crush you overnight 
→ Liability when AI screws up

𝐀𝐈 𝐞𝐭𝐡𝐢𝐜𝐬 𝐢𝐬 𝐭𝐡𝐞 𝐜𝐨𝐦𝐩𝐚𝐬𝐬.

Ethics isn't there to slow us down. It's what will separate businesses that scale responsibly from those that implode spectacularly when their automated systems go rogue.

AI for the people, not against them.

The AI skill making the biggest comeback of the year is... 𝘱𝘳𝘰𝘮𝘱𝘵𝘪𝘯𝘨.

Once a must (GPT-3 times), then gradually a hype (GPT-4 times), prompting is back stronger than ever with GPT-5 (or you will have a very, very painful experience).

And I'm not here to debate whether we will still need prompting skills 10 years from now, because I can't predict the future.

I can only tell you what you need in the present. 

Everything in AI today comes down to text.

Therefore, you would expect that some text inputs perform better than others. 

And that learning how to write better inputs makes, well, a ton of sense.

So yeah, prompt engineering is back. It never really left.

Come learn it with us tomorrow:

𝐏𝐫𝐨𝐦𝐩𝐭 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐢𝐧𝐠 𝐟𝐨𝐫 𝐁𝐞𝐠𝐢𝐧𝐧𝐞𝐫𝐬
👤 𝐇𝐨𝐬𝐭: Jan Meinecke
🕒 𝐓𝐢𝐦𝐞: Wednesday, August 20th, 2025, 5:00pm CET
📍 𝐋𝐨𝐜𝐚𝐭𝐢𝐨𝐧: Online

Want to register? Let me know in the comments ✌️

Why spend the summer building sandcastles… when you could be building 𝘢𝘶𝘵𝘰𝘮𝘢𝘵𝘪𝘰𝘯𝘴?

The 𝐀𝐈 𝐁𝐮𝐢𝐥𝐝𝐞𝐫 course is €699 all summer long ☀️

And it's a little bit more than a course...  it’s a 𝘧𝘶𝘭𝘭-𝘢𝘤𝘤𝘦𝘴𝘴 𝘱𝘢𝘴𝘴 to everything we create at 9x:

→ 25+ 𝐨𝐧-𝐝𝐞𝐦𝐚𝐧𝐝 𝐥𝐞𝐬𝐬𝐨𝐧𝐬
Learn to build and ship real AI automations from scratch — no fluff, no theory marathons.

→ 𝐓𝐨𝐨𝐥 𝐝𝐞𝐦𝐨𝐬
Step-by-step walkthroughs of the exact AI tools we use to automate workflows.

→ 𝐇𝐚𝐧𝐝𝐬-𝐨𝐧 𝐞𝐱𝐞𝐫𝐜𝐢𝐬𝐞𝐬
Apply what you learn immediately. Build, test, and deploy as you go.

→ 𝐏𝐫𝐢𝐯𝐚𝐭𝐞 𝐜𝐨𝐦𝐦𝐮𝐧𝐢𝐭𝐲 𝐚𝐜𝐜𝐞𝐬𝐬
Get direct answers from us, ask questions, share wins, and learn from other operators.

→ 50+ 𝐫𝐞𝐜𝐨𝐫𝐝𝐞𝐝 𝐥𝐢𝐯𝐞 𝐰𝐨𝐫𝐤𝐬𝐡𝐨𝐩𝐬
Deep dives into real-world use cases, new tools, and practical AI strategies. New sessions added each week.

→ 𝐅𝐮𝐥𝐥 𝐚𝐜𝐜𝐞𝐬𝐬 𝐭𝐨 𝐭𝐡𝐞 𝐀𝐈 𝐎𝐩𝐞𝐫𝐚𝐭𝐨𝐫 𝐜𝐨𝐮𝐫𝐬𝐞
10 focused lessons on using ChatGPT, Claude, and Gemini to work faster and smarter.

→ 𝐌𝐨𝐧𝐭𝐡𝐥𝐲 𝐮𝐩𝐝𝐚𝐭𝐞𝐬
Stay ahead. The course evolves with AI—new tools, new workflows, always current.

Offer ends August 31. 

Clock’s ticking!

Going fishing tomorrow. Here are 10 lessons fishing taught me about B2B SaaS Sales:

I usually enjoy explaining complex stuff, and I am pretty good at it most times. But today, the EU AI Act won.

7️⃣0️⃣7️⃣ registrations for tomorrow's webinar... don't miss this plane! And get ready to take-off with n8n agents:

𝐁𝐮𝐢𝐥𝐝 𝐚𝐧 𝐧𝟖𝐧 𝐀𝐈 𝐀𝐠𝐞𝐧𝐭 𝐥𝐢𝐯𝐞 𝐰𝐢𝐭𝐡 𝐦𝐞
👤 𝐇𝐨𝐬𝐭: Jan Meinecke
🕒 𝐓𝐢𝐦𝐞: Wednesday, August 27th, 2025
📍 𝐋𝐨𝐜𝐚𝐭𝐢𝐨𝐧: Online

Comment if you still haven't registered and I'll share with you the event link.

This is 100% what separates companies from unlocking massive gains from AI today.

The models can get 10x or even 100x better it won‘t matter if you don‘t feed them your context.

For this to happen you need to:
1. Pick and implement the right tools
2. Orchestrate them 

Some SaaS like Super will massively help you get closer…

… and for the last mile you will need to build custom automations.

Which is why learning to build with AI is going to quickly become the most valuable skill - irrespective of your current role.

After 100+ calls with companies on how they use AI, I found 2 main productivity unlocks left.
Most are missing both...

1. Orchestration

2. Context

This is where the real competitive advantage lies. And only the best AI native companies talk about this right now.

The data context gains are invisible, but my god, they're mindblowing. When you analyze a sales demo:
- Generic AI response: "This was a positive call with good engagement. The prospect seems interested."
- Context-aware AI response: "Based on 47 similar Enterprise demos, this prospect shows the objection pattern of our highest-converting deals. They mentioned 'compliance requirements' 3 times - same language as Acme Corp before they signed. Recommend highlighting SOC2 certification and scheduling technical deep-dive within 72 hours."

One gives you a pat on the head. The other gives you a playbook.

The AI models themselves are commoditized. Prompting as well. Hence, so are their gains.

What's not commoditized, is having your company's knowledge available to any AI, anywhere, instantly.

Your sales conversations contain decision patterns. Your product feedback reveals what actually drives adoption. Your support tickets show exactly what makes customers stick around or leave.

The real challenge? Getting current company context to your AI tools when that context changes every single day. New demos happen. Fresh feedback arrives. Recent tickets get resolved.

Static knowledge bases become outdated the moment you create them.

We built Super to solve this.
Use it as an MCP to bring your team knowledge in any tool or model you want to use.

Whether you're using Claude for strategy analysis, chatGPT for writing, or whatever product launches next month, Super delivers your most current company context instantly.

Your competitors can copy your AI features in weeks. They can't replicate years of integrated, always-updated company knowledge flowing to every AI interaction.

The companies winning at AI aren't those with access to the best models.
They're the ones with the best context delivery system.

P.S. Beyond the app, Super works with any AI tool through its MCP. 
Check it out if you want your AI to actually understand how your business works.

Taplio

Alexandre Kantjas's Linkedin Analytics

What is Alexandre talking about?

Who is engaging with Alexandre

's Best Posts (last 30 days)

Want to drive more opportunities from LinkedIn?