If you’re a founder or business leader interested in AI, your social feed is full of influencers demonstrating their no-code websites and workflow automation projects on YouTube and Twitter/X.

You’ve either dreamed about or started experimenting with ways to save your team’s time using these no-code platforms.

Is it sensible to build on these platforms? How do you choose the right one for you and your organization? What are the pitfalls?

In this post, we will answer these questions.

A brief history of no-code

Where is this no-code trend coming from?

No code workflow automation is not new. It started with spreadsheets, especially Microsoft Excel, which was launched in 1985. Excel is one of the most widely adopted business software on the planet, with around 1B monthly users.

The rise of APIs in the 2010s led to an explosion of no and low-code workflow automation platforms with better connectivity to business applications. Over time, the space consolidated around a few leading competitors in each market segment, such as Airtable, Zapier, and Make for SMEs (small and mid-sized enterprises), and Microsoft, Salesforce, and ServiceNow for large enterprises.

In the late 2010s, the no-code space for front-end interfaces grew as well and matured with a few leading companies: Shopify, Wix, Squarespace, and Webflow in the SME segment, and Microsoft, Salesforce, OutSystems, and SAP in the large enterprise segment.

Today, no-code is a $30B/year business growing at 25%/year.

Microsoft and Salesforce lead the market with multi-billion dollar workflow automation businesses. ServiceNow makes around $3 billion annually, while OutSystems makes around $500-800 million. Shopify, Wix, and Squarespace are in the $1-2 billion range. Zapier, Dataiku, Make, and Airtable are estimated to make $300-400 million each.

However, a new generation of AI-first no-code platforms is emerging, aiming to offer unique functionalities to business users.

For workflow automation, you’ve heard of n8n, Pipedream, Gumloop, Lindy, and Retool. For front-end/app development, the most talked-about startups include Bolt.new, Lovable, Softr, UXPilot. (There is some overlap between these categories).

With this context in mind, let’s discuss what you need to know before choosing a no-code AI automation platform.

1. No-code is getting better, but you should definitely consider code

No-code platforms are leverarging AI to dramatically improve user experience.

Specifically:

1. GenAI can significantly speed up app creation. You write what you want, a chatbot builds it, and then move some boxes to refine your creation.

2. GenAI enhances app capabilities by introducing innovative functionalities such as AI agents understanding natural language and deciding actions autonomously.

At the moment, #1 is hit or miss, but #2 works well.

However, at the same time, coding is becoming easier. Applications like Cursor, Github Copilot, Windsurf, Cline, and Replit have introduced agents that generate code iteratively based on natural language instructions.

This raises the question: shouldn’t you just code instead of using a no-code platform?

If coding isn’t too intimidating for you, now is a great time to start. Consider coding instead of a no-code platform subscription.

2. If you are going for no or low-code, prefer low-code

A platform that supports code editing offers far more customization possibilities than one focused on visual app design, given that code editing is easier than ever. If you are going to choose no or low-code, prefer low-code.

If you only use no-code platforms, it’s likely that a single platform won’t meet all your needs and you’ll need to switch to implement various AI automations.

Among workflow automation platforms, Pipedream supports custom code well. n8n and Lindy have launched it in beta and need to enhance the feature.

In front-end development, Bubble, Replit, Lovable, Bolt.new and Replit are pretty good at supporting customization.

Two caveats:

• Ironically, low-code platforms currently struggle with AI-powered autocomplete in their code interfaces. To generate code, you may need to open Cursor or another app, chat with the agent, and then copy and paste the code into the low-code app. Hopefully, the user experience will improve soon.

• While code generation is easy, hosting a live app can be technically challenging. Platforms like Replit, Lovable, and Bolt.new are effective at bridging the gap between code and app deployment.

3. Make sure you understand how app usage drives subscription costs

No and low-code subscription costs can escalate quickly due to pricing models based on user count and the number of monthly actions.

For example, Softr charges front-end apps per end-user. This makes subscription costs more suited to internal company applications than external-facing ones.

As for n8n (and many other workflow automation apps), they charge by number of executions, with self-service pricing tiers limited to 10,000 monthly tasks.

If you’re building an app that automates a lot of low-value interactions, you need to consider costs carefully. For example, an app that manages hundreds or thousands of social media influencers, or an app used by your company’s vendors to submit their bills for automated processing.

In a scenario with many users and/or monthly tasks, selecting Bubble for the front-end and Replit for the back-end may be the right approach, given their pricing.

However, Bubble and Replit don’t offer the security and reliability guarantees that larger enterprises expect. If your requirements are stringent, be prepared to pay per user.

4. Compare multiple platforms to pinpoint their strengths and limitations

Each no/low-code platform represents a functionality trade-off along the following dimensions:

To understand the differentiating features of each platform, you need to ask the following questions:

• Is the platform primarily for SMEs/startups or large enterprises?

• Is the platform more focused on UI creation (front-end) or on workflow automation (back-end)?

• Tech skills required: What level of coding skills do you need to use the platform properly (no / beginner / intermediate)?

• Code customization: Can you insert or edit code to customize the automation?

• AI workflows: Does the platform facilitate the deployment of AI assistants to execute workflows and call external services in a no/low-code way?

• App creation chatbot: Is it possible to describe your requirements in natural language to get the platform to build your app?

• Pricing model: Is the platform priced by project, by user, or by task?

Let’s compare a few of the most talked-about startups: