From Zero To SaaS #42 Building AI MVPs

The cost and speed of going to market with an application have never been this low, but is that working in your favor or against you?

Competition can force innovation or destruction, depending on how you play the business game.

In pre-generative AI, the lean start-up strategy was the most adaptable for decreasing the risk of failure.

Although the lead start-up has helped thousands of companies achieve product-market fit, it is vulnerable to bias and uses compartmentalization.

The newer trend, Build In Public, has changed the game.

Instead of de-risking by doing many growth experiments, you build a community around your product, which naturally attracts innovators and early adopters.

In addition to the insight they provide, these people will follow you to the next product, even if you fail.

Because of the parasocial relationship, you spend more than 7 hours with someone's content, and you are much more lenient when they make mistakes because you have compassion for their journey.

Build From Scratch v.s using Platform Builder

No Code

In most MVPs, you would be faster in a building using a third-party platform, like Bubble or Webflow, which does the hosting, and you only take care of API integrations and design.

Most platforms have connections to automation providers like Make, Airtable, etc., to make the niche API calls availble.

However, this is a short-term solution because you are vulnerable to the whims of third-party providers who can change prices, remove functionality, and go out of business, and their priority is never your business.

Low Code

Nowadays, apps such as Lovable and Bolt can design and code the front end of an app with just prompting, but the back end of the API connections can be more challenging.

If you choose this route, you should first design the app's architecture with ChatGPT so you will know how to prompt these models.

Most SaaS only do web scraping (Gaining Information) and Automation (Action Taking With That Information).

• Apify: Give you access to thousands of premade API call scripts to plugin
• NPM: Gives you access to thousands of Javascript frameworks/libraries
• Clay API: Can do the entire enrichment and automation for B2B companies
• Airtable API: Can do all the database operations for you
• Relevance AI API: Can Add Agentic Automation to your platform

High Code

If you build it with code, use Javascript instead of Python because Chrome is built for Javascript.

I use Vite to have the fastest Front End.

Any cloud hosting provider is acceptable. However, I prefer Google Cloud over AWS not because they are the cheapest but because Google is constantly building innovative products on top of its cloud.

However, if you are a fan of Microsoft, you should choose Azure because you will be more familiar with their communication style.

If you don't want path dependency, go for the backend with Supobase.

Outsourcing Design and Branding

Take it from me, who has lost a couple of thousand on design:

Design isn't essential in the early stages.

Innovators are happy if it works, unlike the early majority, who choose based on how easy it is to use. They are also not price-sensitive like the late majority.

BUT

Without Speed-To-Value, nobody will use the 'early stage' product.

The people annoyed enough to look for a solution online don't want to add much learning of how to use products to get what they want.

So I use:

• KISS = Keep It Simple Stupid; a teenager should be able to use your platform without industry-specific knowledge.
• Templated Design = Go to Dribble, find how the UI looks in your industry, and copy best practices. Even if you are going to hire a designer, please do this because it is almost impossible to talk about design without showing visual examples.
• Design Systems = An example is Google's Material Design, which has collections of fonts and icons that work well, so you don't have to reinvent the wheel. You can find these on Figma Community Templates.
• Design Sprint = Developed by Google to build apps within one week instead of months.

As a measuring stick, you can get an excellent design for under 2k on Fiverr or hire a second to third-year freelancer.

I don't recommend hiring a design agency because most startups don't have a logo or brand identity.

A design sprint teaches more about how your cofounder or client views the product than creating the business canvas together.

Taking Care Of Legal

Tight Lips

Since I believe in Win-Win markets, I am not focused on this area.

But for everyone walking around with an AI idea, don't think you are the only one.

The higher you go up in the chain, the more NDAs will become common.

Another trend is just building in Stealth Mode.

Eradicating Future Competition

Look at the Facebook story, and you understand why the non-compete clause exists.

However, there is a better way to ensure that you and your co-founder are on the same page in the short, medium, and long term.

You need to align not with the product but with what doing business means, especially when the market keeps changing and people are entering different life stages.

For example, in my last call with Nick, we bonded over our shared love of the Poppins font. These small things make business partnerships work.

If you want to be updated on the Content Design Tool ?? https://nick-broekema.kit.com/ed3b81cbd9

Loyalty

Likewise, if you take the time to invest in your employees who see for themselves that they become better each year, even if they leave, they won't hurt the company.

What goes wrong in most start-ups and scale-ups is that they talk about promoting someone but do not put money where their mouth is.

Yes, you take risks by putting someone into a managing position without experience, but experience is a matter of time.

Loyalty is not something you buy; it is something you gain by giving more to your customers than financial compensation.

Adapting to Legislation

For AI companies, we have three problematic cases.

1. There is no legislation on AI-generated data.
2. If web scraping becomes illegal, companies that have done it in the past will have too big an advantage.
3. Agency problem: If an agent comes up with the idea to web scrape or automate a task, but a human never specifies that, who is in the 'wrong'?
4. Humancentric problem: What are the rights and responsibilities of agents online and, for example, future robots in the workplace or house? If you decide that, is that not super biased towards on-form existence?

This is how I see the world; I come from web scraping in the B2B industry.

GDPR is the biggest joke in the world because it is tough to prove a company used someone's data, and it is costly to go to court.

For example, LinkedIn sued a company to end web scraping on their platform, and they lost.

They try to counter this in the user agreement, but the latest court decision overrules this.

I often reflect on how disruptive technology like quantum computing, AI models, robots, and other technological advancements will impact ownership discourse.

If we look at the abundance of reality, not just the earth, I have the position that we will fade out money exists nowadays because it builds a perspective of scarcity and that stops existing:

• Bio Bioengineering can decrease the speed of aging and, in some situations, reverse the cellular age of specific organs.
• 3D printing would eventually be able to print food and materials abundantly and remove the need to move all these resources worldwide.
• Look at YouTube, which has Stanford courses; education is neither class nor location-dependent.
• Microsoft's latest quantum computing chips, which use a new form of matter, bring quantum chips to small electronic devices, making the computing power even cheaper.

There still will be differences in health, wealth, etc, because humanity doesn't adapt to the masses quickly.

But I don't see a system where banks, governments, and corporate interests determine the course of humanity like they do nowadays.