Learn the four types of AI builders and why one wins big

AI with ALLIE

The professional’s guide to quick AI bites for your personal life, work life, and beyond.

In today's newsletter, I'm breaking down the four types of AI builders, why the fourth category has tremendous potential, and exactly how you can join them if this resonates with you.

Are you languishing with AI?

Two and a half years into the generative AI revolution, and some users are still stuck. I’ve met hundreds of professionals who feel like the AI tools they’re using are either "too simple" (read: not powerful enough) and "too technical", or simply that the rest of the world has cracked the code to using AI tools and they just haven’t been handed the secret playbook.

It’s becoming more and more obvious to me: there’s a new category of AI builder that is absolutely crushing it right now.

If you like crossword puzzles, or spreadsheets, or spy movies, or rube goldberg machines, or escape rooms, or those wooden boxes that you can’t open from the top like a normal person but that you have to slide out some secret key from the side (why do I love those things so much?), you’re going to want to read on.

The Three Builders Most People Know

Software (including AI tools) sorta breaks up into three categories:

No-code tools like ChatGPT or Otter AI or PowerPoint where you can just type and click around and type things in or speak to make things work. No coding required. The input is incredibly intuitive and billions of people can use it without breaking a sweat.

Low-code platforms like n8n, where you may need a bit of code to put something together, but it's a light lift for a technical person. It's a few scripts here, a GitHub integration there. But it's targeted at software engineers, data scientists, and analysts. Very STEM-oriented in its interface. The audience is maybe 50-100M people.

Pro-code environments like Amazon Bedrock, Databricks, or HuggingFace, where it's expected that engineers will code and build, sometimes from scratch. The audience for these is less than 30 million globally.

So far, most things have fallen into these three buckets. And until now, people generally did too: if you had no coding experience, you stuck with no-code tools; if you were more technical, you could take advantage of low-code or pro-code solutions. Builders stayed in their building lanes.

But There's a Fourth Category

There is a fourth category that hasn't been talked about much but is getting a ton of AI value right now.

These are people who aren't staying in their lane. They have more systems-thinking and technical intuition than typical no-code users. They have higher resilience and grit. They have more of a hacker mindset. But they can't and don't want to learn how to code. They know they need to automate and customize workflows but prefer to stay away from traditional coding environments.

That is: the Composer.

(I’ll explain why this isn’t exactly a “Vibe Coder” but for the purposes of this email, you can use those terms interchangeably.)

These builders don't fit the traditional categories, and neither do the tools they gravitate toward. They're using platforms like Gumloop, Copilot Flows, Airtable, Base44, Clay AI, and even Claude Artifacts—tools that are more sophisticated than typical no-code solutions but don't require actual programming skills. These tools require you to think more about data flows, conditional logic, and integrations, but they let you build through visual interfaces and configuration rather than just code.

Let me show you some examples…

They're both category breakers—builders and tools that refuse to be boxed in.

And that’s even more critical today because simple chat thread interfaces are dying. Actually: chat is dead. You need to prepare for the messy interface shift, toward more complicated systems, flowcharts, spreadsheets, and vibe coding.

Why Scrappier Builders Are Winning (And What "Scrappy" Actually Means)

Let me be clear about what I mean by "scrappy." I mean resourceful. Determined. Willing to figure things out through trial and error rather than waiting for the perfect, polished solution. It's about having that hacker mindset without needing to write code.

Scrappy Composers see a complex tool and think "challenge accepted" instead of "this is too hard." They're comfortable with a steeper learning curve if it means they can build something more powerful. They don't need everything handed to them on a silver platter. They'll dig in and figure out how to make it work.

And notably, they know they can use AI in this challenge too! I spoke with a Composer last week who has ever written a single line of code in her life. She wanted to set up a workflow automation between Google Slides, Google Sheets, ChatGPT, a CRM, and webhooks to fully automate sales pitch deck creation. And she did it with Claude 4 Sonnet in one window and her workspace in another. Every single step of the way, she went to Claude. What’s a webhook? Claude. It’s giving me this weird error that says I need to import a new library? Claude. Where do I find extensions in Google Sheets? Claude. How can I improve the font and layout? Claude. How should I best set up this prompt? Claude. She showed off an entire JSON structure she built with Claude to improve the reliability of her integrations. She’s a good friend who reads this newsletter religiously. To her I say, hell yes.

The companies building these more complex tools actually need to lean into this new builder, because they're a big winner in the AI age.

They're willing to watch a 2-3 hour tutorial to learn a tool. They're willing to stare at an uglier UI to get big value. If I were you, and I were a complete non-coder and I thought I had the tinkerer's urge in me, I would spend some time leveling up to this builder type, and not learning to code to be able to use low-code tools.

Are You a Scrappy Composer?

Here's how to know if you've got the scrappy builder mindset:

• You've spent an entire weekend figuring out how to make something work, just because you were curious

• You don't mind if a tool looks a little rough around the edges, as long as it's valuable

• You prefer to understand how something works rather than just use it

• You're comfortable with trial and error as a learning method

• You get excited by the idea of building custom solutions for your specific problems

If that sounds like you, welcome to the Scrappy Squad.

What This Means For You

Here are 5 things you can do today to embrace your inner scrappy Composer:

1. Explore scrappy tools and find your fit

   Start by trying a few different types. Maybe a flowchart builder like Gumloop, a data tool like Clay AI, or an advanced spreadsheet like Airtable. Spend a weekend playing with 2-3 options to see which interface and problem-solving approach clicks with your brain.

2. Budget real time for the learning curve

   Block out some focused hours. Maybe a Tuesday evening or your lunch break for a week to work through tutorials and documentation. Get snacks ready and get yourself psyched for the journey of watching stuff constantly break.

3. Start with someone else's template, then make it yours

   Browse community galleries or template libraries within your chosen tool. Most platforms like Airtable, Clay, or Gumloop have them (the Make.com ones are incredibly easy). Find a workflow that's solving a similar problem to yours, pr part of it, duplicate it, make sure it runs, then start tweaking. Change the data sources, modify the logic, add your own steps. Learn the platform's capabilities while building something useful.

4. Embrace the inevitable messiness

   Your first few builds will be inefficient, confusing, or downright ugly. When you know that going in, you will feel less like an idiot when it happens. You are not a failure, ok? You are not a failure. That's the learning process. Keep iterating until it works. There's something deeply satisfying about finally getting something to work when YOU made it happen, especially after wrestling through the clunky-murky-almost-giving-up stage.

5. Document and share your discoveries

   Keep notes on what worked, what didn't, and the creative workarounds you discovered. Then share them in tool communities, on LinkedIn, or with colleagues. Try something like Scribe, Loom, Jumpshare, or OBX to record a quick demo.

The Bottom Line

The AI landscape is creating space for a new type of builder, a scrappy systems-builder who thrives on slightly more complex interfaces, something between simple tools and engineering platforms. If you've ever felt too curious for basic automation but don't want to become a full-time programmer, this is your moment.

Roll up your sleeves and make Claude explain every little last detail to you. There is no wall you can’t break through.

Stay curious, stay scrappy,

Allie

INTRODUCING: The AI-First Conference

I get asked all the time:
“How do I actually build an AI-First business or career?”

Now I have an answer: Join the AI-First Conference this August. 

I’ve officially launched the AI-First Conference, a virtual experience designed to help you stop using AI for small gains and start using it to fundamentally rethink how you work, build, and compete.

You’ll walk away with:

✅ An AI-First mindset to help you fundamentally rethink how you work and lead
✅ Concrete workflows to implement across your business or team
✅ The 4 modes of working with AI (beyond basic prompting)
✅ Playbooks and use cases to implement immediately
✅ New mental models to compete with and win—even if your team or business is small

If you’ve been dabbling with AI on the side, this is how you stop dabbling and start competing—with clarity.

Feedback is a Gift

I would love to know what you thought of this newsletter and any feedback you have for me. Do you have a favorite part? Wish I would change something? Felt confused? Please reply and share your thoughts or just take the poll below so I can continue to improve and deliver value for you all.