Hi, Sparkbuddies!
Okay, real talk: if you're not using AI in 2025, you're basically showing up to a Formula 1 race on a bicycle. Since ChatGPT dropped and absolutely broke the internet in late 2022, generative AI has become the ultimate career cheat code. And if you're running a startup, nonprofit, or literally any venture, you NEED to get on this train before it leaves the station.
This guide is for all my non-tech founders out there who want to level up without learning to code.
What is generative AI?
Think of generative AI (GenAI for short) as your super-smart creative assistant that can pump out text, images, and sounds based on whatever you throw at it. It's basically trained to think "What would a human create here?" and then... creates it.
It’s not the only type of AI system out there: classification systems (e.g., Gmail’s spam detector), regression systems (e.g., a bank’s credit default prediction engine), recommendation systems (e.g., Netflix’s movie recommendations), computer vision (e.g., Tesla’s self-driving system) are also AI systems.
However, generative AI is the one type of AI system that has experienced the largest investments and greatest progress during the last few years, thanks to the invention of a new machine learning method (the Transformers architecture), the online availability of most of human knowledge, and the broad availability of powerful training chips (GPUs).
Here’s what happens when you ask ChatGPT, Anthropic’s Claude or Google’s Gemini to respond to your instructions:
Your instructions, together with hidden system instructions and the chat history, are converted into a sequence of numbers (called tokens).
The large language model (LLM) predicts a number (token) representing the first word in how a human would likely respond to these instructions, based on training data. Prediction takes place iteratively, token by token, where each new token is incorporated into the chat history for the prediction of the next token.
The numerical output is converted back into text, image or sound, and displayed in the user interface.
Pretty wild, right?
What can generative AI be used for?
Here's where it gets exciting: generative AI is basically automating the knowledge economy. Translation? Those tasks that used to eat up your entire weekend can now be done in minutes.
AI tools can:
Understand pretty much anything you throw at them (emails, documents, contracts, your rambling voice memos, images)
Create whatever you need (blog posts, code, spreadsheets, podcasts, videos, you name it)
Take action online (search databases, connect with apps, navigate websites)
Think of AI as that incredibly capable intern who never gets tired, never asks for coffee breaks, and actually follows instructions (as long as you're clear about what you want).
How to become fluent in AI?
At Sparkbuddies, we recommend a 5-step process so that non-technical founders become fluent with AI tools at every level of their entrepreneurial journey:
Use chatbots to assist you in your online tasks.
Stay up to date on how others are using AI in their day-to-day work.
Use specialized AI-powered tools to increase your personal productivity.
Create agents to automate repetitive tasks.
Experiment with coding assistants to fully experience the power of software written in natural language.
Let’s review each of these steps one by one.
Quick heads up: You'll need to sign up for various platforms along the way. Most have free trials, but expect to spend around 20 USD/month for the premium features. Totally worth it, and you can always cancel after testing things out.
Step 1. Use chatbots to assist you in your online tasks
Let's start with the big five:
ChatGPT (chatgpt.com): The OG that started it all. Incredibly versatile, though sometimes it can be a bit... overly enthusiastic with its responses.
Google Gemini (gemini.google.com): Super powerful and plays really nice with Google's ecosystem, but it's pretty cautious with its responses.
Anthropic Claude (claude.ai): My personal favorite for longer projects and anything involving code. It's like having a really thoughtful research assistant.
Perplexity (perplexity.ai): The search genius. It gives you answers WITH sources, so you can actually verify what it's telling you.
Microsoft Copilot: Lives inside Microsoft apps. Honestly, it's the weakest link, but worth knowing about.
These chatbots offer mobile apps too.
You should use one or several of them every day and experiment with the various model options offered by each to understand their strengths and limitations. The more advanced the model, the better its reasoning capabilities, but also the slower it gets.
For example:
Need to write an email? Ask the chatbot to review and improve it.
Need to prepare for an interview? Ask the chatbot to suggest questions and answers.
Need to create an illustration for a presentation? Ask the chatbot to generate it.
Need to do a web search? Ask the chatbot to do it for you.
Need to post on social media? Ask the chatbot to propose ideas.
Need to summarize or translate a text or a PDF? Ask the chatbot.
Need to cook veal marsala? Ask the chatbot for the recipe.
Etc.
The more you experiment, the better you'll get at giving clear instructions. And trust me, clear instructions = mind-blowing results.
Step 2. Stay up to date on how others are using AI in their day-to-day work
New AI tools drop literally every day. The best way to stay current? YouTube is your best friend.
Thousands of creators are testing, reviewing, and demonstrating the latest tools. Start with these searches and watch your feed transform into an AI goldmine:
"AI image generation tutorials"
"Chat with PDFs using AI"
"AI productivity hacks"
"Generative AI for business"
"What is RAG explained"
Once YouTube's algorithm figures out you're into this stuff, you'll have endless content recommendations.
Step 3. Use specialized AI-powered tools to increase your personal productivity
This is where things get really fun. Tons of startups are building user-friendly apps that make AI accessible to everyone. Most of these are basically fancy interfaces for the same AI models from OpenAI, Google and Anthropic, but that's exactly what makes them so useful!
Here are the main use cases that you should definitely experiment with:
Meeting recording, transcription and summarization. Possible tools include Notion, Granola, Otter, Fireflies, Fathom, Zoom, Google Meet… We recommend starting with Notion’s AI meeting notes.
Image generation. Possible tools include MidJourney, ChatGPT, Gemini, Flux1, Stockimg, Photoroom, Ideogram, Leonardo… We recommend starting with ChatGPT for image generation and Google Gemini for image editing.
Video generation. Possible tools include Kling, Argil, HeyGen, Captions.AI, Viggle, Synthesia, Descript, Hailuo, Pika, Runway, Veo 3… You can get started with Runway, but you will probably need to use several tools to get to the desired results. Watch YouTube videos to see examples of workflows.
Voice generation. Possible tools include Assembly AI, ElevenLabs, Ultravox, OpenAI… We recommend starting with ElevenLabs.
To learn about these tools and discover new ones, search for tutorials on YouTube depending on what you are trying to do.
Step 4. Create agents to automate repetitive tasks
Ready for the advanced moves? Time to automate those repetitive tasks that are eating your life.
Think about what you do manually every week. Here are some workflows our team has automated:
Scan all invoices received by a specific email address and put their details into an Airtable spreadsheet.
Respond to community emails, leveraging our knowledge base.
Transcribe and summarize handwritten notes added as PDFs to a specific folder in Dropbox.
Search for young startup founders and find their contact information online to propose an interview.
Etc.
There are two main types of tools to automate these types of workflows:
Automation platforms. Those are no-code platforms where you can program workflows. Possible tools include Zapier, Make, n8n, Gumloop, Lindy, … To get started, we recommend Lindy.
These tools have a somewhat steep learning curve. To experiment with them, search for tutorials on YouTube for your specific workflow.
Step 5. Experiment with coding assistants to fully experience the power of software written in natural language
The promise of AI is that anyone can become an app developer, thanks to GenAI tools that convert natural language instructions into code.
The reality is that we are not there yet. Coding assistants are powerful, but they require precise instructions to be effective, and you can get stuck in development dead-ends if you know nothing about app development.
Still, these tools are mature enough that anyone can experiment with them to get a start on their AI journey. We recommend experimenting as much as you can.
For non-technical people, app development tools include Lovable, Bolt, and v0. We recommend starting with Lovable. Think of a website or an app that you’d like to create, and see how far you can get by giving natural language instructions to Lovable’s chatbot.
Even if you get stuck, you will get ahead simply by being an early adopter and getting ready to become a vibe developer as these tools get better.
For a deeper perspective on this topic, watch Andrej Karpathy’s video describing how natural language is the new programming language:
What you’ve learned
You've just learned the blueprint for AI fluency:
✅ Master AI chatbots through smart prompting
✅ Discover cutting-edge tools via YouTube
✅ Boost productivity with specialized AI apps
✅ Automate repetitive tasks using AI agents
✅ Experiment with natural language programming
Here's the truth: AI fluency isn't a destination, it's a journey. Start small, practice consistently, and keep expanding as new tools emerge.
It’s also hard work. Let’s be honest, if you are competing with AI-savvy founders or professionals, becoming fluent in AI tools is key to staying relevant.