5 Unexpected Lessons from Teaching 20,000 People About AI

Google invited me to teach entrepreneurs about AI. What followed was an interesting sequence of realizations and unexpected learnings.

"So, you're like... an AI expert?"

That's what someone asked me before my talk at Google's Skills of Tomorrow AI course. I laughed. The truth is, nobody's really an "AI expert" these days - we're all learning as we go. What I am is a product guy who's been in the trenches, implementing AI at companies like Allegro, Booking.com, and now GOG.

But here's the thing: speaking at this program taught me way more than I expected. With 20,000 participants from all walks of life, it was like getting a crash course in how AI is actually being adopted (or not) across different industries. Let me share the real, unfiltered lessons I learned.

1. The Truth About AI Adoption in Poland (It's Not Pretty, But There's Hope)

Let's cut to the chase: Poland is lagging behind in AI adoption. We're hanging out near the bottom of the EU rankings with Bulgaria and Romania. This might shock you, especially if you've seen how tech-savvy Polish companies can be. I mean, our banking apps make German ones look like they're from the 90s, and everyone here uses BLIK while other countries are still figuring out digital payments.

The Raw Numbers (Brace Yourself)

According to GUS:

• 94.1% of Polish companies haven't touched AI

• 3.2% are thinking about it (but not doing anything)

• Only 1.9% cite lack of expertise as the main barrier

• 1.7% blame high implementation costs

• A tiny 0.9% worry about ethical concerns

💡 Reality Check: Think your company can afford to wait? Here's a fun fact: the Ministry of Digitization estimates proper AI implementation could boost Polish GDP by 8% in a decade. That's billions of złoty on the table. Some of it could be yours.

What Can We Do About It?

The 20,000 AI leaders emerging from the Skills of Tomorrow program are just a drop in the ocean of what's needed. That's why it's crucial for each graduate to become an AI ambassador within their organization. Here are some practical steps:

1. Start Small, Think Big

   • Begin with simple AI implementations in your daily work

   • Document successes and failures

   • Share learnings with your team

2. Build Internal Capacity

   • Organize AI awareness sessions

   • Create a knowledge-sharing platform

   • Identify AI champions in different departments

3. Create Quick Wins

   • Focus on low-hanging fruit first

   • Measure and communicate results

   • Use successes to build momentum

2. The AI Interest Spectrum is Massive (And Growing!)

A Diverse Audience

The Skills of Tomorrow participants represented an incredibly diverse professional landscape:

• HR professionals

• Small and medium business owners

• Artists

• Students

• Developers

• Content creators

• Lawyers

• Cybersecurity experts

• And many more

This diversity signals something profound: AI has moved far beyond its traditional domain of data science. What was a niche technical field five years ago and a curiosity two years ago is now touching every aspect of our lives.

The Democratization of AI

The accessibility of AI tools is increasing at an unprecedented rate. Until recently, running a language model locally was beyond the reach of average consumers with smartphones. Now, thanks to model distillation techniques like those used in Deepseek R1, we're approaching a future where everyone can have their own personal, secure AI assistant.

The Implications Are Enormous

This democratization creates vast opportunities across nearly every field. Consider these emerging trends:

1. Content Creation

   • AI-assisted writing and editing

   • Image and video generation

   • Music composition and sound design

2. Business Operations

   • Automated customer service

   • Inventory management

   • Market analysis and forecasting

3. Professional Services

   • Legal document analysis

   • Medical diagnosis assistance

   • Financial planning and analysis

4. Personal Productivity

   • Email management

   • Schedule optimization

   • Research assistance

Why This Matters

The widespread interest in AI across professions indicates we're at a tipping point. The question is no longer whether AI will impact your field, but how quickly and dramatically it will transform it. That's why, regardless of your profession, the best time to start engaging with AI was yesterday. The second-best time is now.

This revolution will require continuous learning from all of us. It's one of the main reasons I started this blog – to help you navigate through the AI hype and find practical, valuable applications for your work and life.

3. The First Steps Are Often the Hardest (But They Don't Have to Be)

Overcoming Impostor Syndrome

Coming to Skills of Tomorrow, I experienced a touch of impostor syndrome - surely there were hundreds of people more competent than me with more valuable knowledge to share. While that's probably true (and why I encourage learning from multiple sources like Andrew Ng, the co-founder of Google Brain, Deeplearning.AI, and Coursera), I discovered that my practical experience from Allegro and Booking.com proved invaluable for those taking their first steps in this field.

Learning from Mistakes: A Real-World Example

Let me share one of my early mistakes as a Product Manager at Allegro. It’s a story about my first AI project - and a perfect example of what NOT to do.

We had this problem with product listings. Sellers would list the same iPhone with different names:

• "iPhone 14 32GB Warsaw"

• "iPhone 14 32 cheap!!!"

• "New iPhone 14 32GB PL"

An example iPhone offer on Allegro with the culprit in the name

Our recommendation system saw these as different products. Our solution? Build a fancy AI system with Siamese neural networks! Sounds impressive, right?

Six months and countless meetings later, we had... nothing useful. Then one developer got fed up, spent a week building a simple text matching script, and boom - problem solved.

Why Smart People Make Dumb AI Decisions

I see this pattern everywhere:

1. Get excited about AI

2. Pick the most complex problem you can find

3. Try to solve it with the fanciest AI solution

4. Fail spectacularly

5. Finally do what you should've done first: start small

🎯 Quick Win Template: Want to avoid this trap? Here's your first AI project template:

1. Open ChatGPT

2. Take one repetitive task you did today (writing emails, summarizing documents, whatever)

3. Try to automate it

That's it. That's your first AI project.

"But what if I mess something up?"
Here's a secret: everyone messes up their first AI projects. The trick is to mess up small and cheap.

4. The AI Toolbox is Vast (And That's Both Good and Bad)

An Overwhelming Landscape

Reviewing the Skills of Tomorrow materials, I was amazed by the variety of AI tools and techniques presented by other speakers. Tools like AION by Franek Gerogiew and NotebookLM represent fantastic AI applications that anyone can start using. I was genuinely fascinated by how many participants began creating their first AI-generated images and videos.

Navigation Strategies

In the flood of new tools and models that promise to "change everything," it's crucial to maintain perspective:

1. Focus on Problems, Not Tools

   • Start with the business problem you're trying to solve

   • Choose tools based on specific needs, not hype

   • Remember that ChatGPT can handle many use cases effectively

2. Evaluation Framework

   • Does it solve a real problem?

   • Is it significantly better than existing solutions?

   • What's the learning curve vs. benefit ratio?

   • How reliable and sustainable is the tool?

3. Implementation Approach

   • Start with proven, mainstream tools

   • Test new tools in controlled environments

   • Build expertise gradually

   • Document what works and what doesn't

Common Pitfalls to Avoid

1. Tool Overload

   • Don't try to use every new AI tool

   • Avoid switching tools frequently

   • Focus on mastering core tools first

2. Feature Fixation

   • Don't get caught up in features you don't need

   • Prioritize reliability over cutting-edge capabilities

   • Consider the total cost of adoption

3. Integration Issues

   • Check compatibility with existing systems

   • Consider data security implications

   • Evaluate maintenance requirements

The Only AI Tools You Actually Need to Start

1. ChatGPT or Claude - for text and basic coding

2. Gemini (or CoPilot if you use Office) - for documents and spreadsheets

3. ONE specialized tool for your industry like image or voice generation

That's it. Really.

⚠️ Warning Signs You're In The Tool Trap:

• You've signed up for more than 3 AI tools this month

• You keep switching tools before mastering any

• You're spending more time learning tools than solving problems

A Day in My AI Life

Here's how I actually use AI:

Morning:

✓ Email or Slack message drafts with Claude 
✓ Meeting summaries with Gemini 
✗ Ignored 5 "revolutionary" new AI tools

Afternoon:

✓ Strategy brainstorming with my "Head of Product" Claude assistant  
✓ Dana analysis using Claude
✗ Resisted urge to try new "game-changing" AI platform

The point? Stick to basics until they don't work anymore.

5. The Human Side (Or Why Speaking to 4,000 People Changed My Mind About AI)

Let me tell you about the most nerve-wracking moment of Skills of Tomorrow: my first live webinar with 4,000 viewers. I was terrified. But something fascinating happened.

The Questions That Changed Everything

People weren't asking about:

• Neural networks

• Machine learning algorithms

• Technical specifications

They asked about:

• "How to convince my board?"

• "How do I start?"

• "What if I make a mistake?"

💡 Key Insight: The biggest barrier to AI adoption isn't technical - it's psychological.

What Actually Works

Based on feedback from AI adopters, here's what helps people get started:

1. The 15-Minute Rule

   • Try any AI tool for just 15 minutes

   • Pick one tiny task

   • Don't worry about "best practices"

   • Just experiment

2. The Copy-Paste Method

   • Find someone who solved a similar problem

   • Copy exactly what they did

   • Adjust only after it works

3. The Public Commitment

   • Tell your team you're trying AI

   • Share your failures

   • Celebrate small wins

What Now? (The Actually Useful Part)

Instead of giving you a generic "AI strategy framework," here's what I want you to do right now:

1. Open ChatGPT (or similar)

2. Paste this prompt:

Task: Help me identify ONE repetitive task in my work that you could help with. My role: [Your Job Title] Main daily activities: [List 3-4 Things You Do Every Day]

3. Try automating that ONE task

4. Reply to this post with what happened

My Challenge to You

Don't plan. Don't strategize. Don't wait for the perfect moment.

Try something today. Fail small. Learn fast.

Because here's what I learned from 20,000 Skills of Tomorrow participants: The people who succeed with AI aren't the smartest or most technical.

They're the ones who started.

PS: The Skills of Tomorrow organization was world-class - huge congratulations to Cezary Jaroni, Michał Domagała, and the entire Google team! Their professional approach to content creation and delivery made a significant difference in how the message was received.

---

Liked what you’ve read? Subscribe to my newsletter or share it with your colleagues. The more feedback I get from you, the better the content will be!