- Growth Case Studies
- Posts
- What Can AI Products Teach Us About Product-Led Growth?
What Can AI Products Teach Us About Product-Led Growth?
Quick product-led growth lessons from ChatGPT, Gemini, Lovable, Replit, Eleven Labs, Canva, and Figma.

In "What is an AI-first product," I outlined a brief timeline of computing's software and hardware history and predicted what AI-first products could look like. Since then, I’ve tried dozens of AI software. As you know, everyone is promising AI tools these days, and everything (recommendation engines, image detection, text prediction, data analysis) is being lumped into this catch-all term.
One thing that’s clear to me is that AI features can enable new product-led growth approaches that drive user acquisition, improve onboarding, activate new users, and even boost retention. We can front-load new user experiences and embed them directly within landing pages, speeding time to value and helping users reach the aha moment faster.
Today’s case study is a rundown of some of these approaches from OpenAI’s ChatGPT, Canva, Lovable, Replit, Google’s Gemini, Figma, and Eleven Labs. For each product, I’ll embed a Loom video. These short videos will not show in email, so please click here to open this essay in your browser to watch them.
This case study focuses on each product’s landing page. A caveat is that I don’t know which marketing campaigns are driving traffic to these pages, how well they’re performing, or what constraints have led to them.
First, a message from this week’s sponsors.
Make Every Platform Work for Your Ads
You’re running an ad.
The same ad. On different platforms. Getting totally different results.
That’s not random: it’s the platform effect.
So stop guessing what works. Understand the bit-sized science behind it.
Join Neuroscientist & Neurons CEO Dr. Thomas Ramsøy for a free on-demand session on how to optimize ads for different platforms.
Register & watch it whenever it fits you.
Startups get Intercom 90% off and Fin AI agent free for 1 year
Join Intercom’s Startup Program to receive a 90% discount, plus Fin free for 1 year.
Get a direct line to your customers with the only complete AI-first customer service solution.
It’s like having a full-time human support agent free for an entire year.
Table of Contents
So, What is Product-Led Growth?
Product-led growth (PLG) is not a new concept. Put simply, it’s a business growth methodology in which the product itself primarily drives user acquisition, activation, retention, and revenue.
We will look at how AI features are driving product-led growth across user acquisition, activation, and retention.
1. OpenAI’s ChatGPT
OpenAI’s ChatGPT chatbot is responsible for the recent AI boom. Since launch, the service has grown to 800 million weekly active users.
ChatGPT’s website puts its product front and centre, letting it do the talking. The hero message “What can I help with?” positions ChatGPT as an assistant, and the text box tells the user to “Ask anything”. In my research, the hero message sometimes changes to “What’s on the agenda today?” “, Ready when you are” and “What’s on your mind today?”
The user can then attach a file (images only for logged-out users), search, study, or talk to the service. This page leaves too much to the prospective user's imagination. I don’t think many people do well with software that presents as a blank canvas. The constraint here however, is that ChatGPT can present literally anything. With infinite possibilities, how does a marketing team decide what use cases to present?
One advantage is that the user can start engaging with ChatGPT without signing up and can immediately begin to see value.
2. Canva AI
Canva AI does a great job of showcasing Canva’s AI capabilities. While this product is product-led growth for onboarding and retention, its landing page is just a showcase. It feels like something’s missing here.
I’d expect Canva to at least allow users to make one request to showcase its capabilities before signing up. Unfortunately, a prospect can only use Canva AI after signing up.
This landing page has its positives. Its hero section showcases how Canva AI works. The GIF shows a user typing their request, selecting a brand kit, and Canva generating three Instagram image options. The rest of the page shows all the features Canva has built into its AI product.
Still, I’d have loved to try this out without signing up. For Canva’s paying users, this is a great way to drive retention and revenue - I have used this feature countless times.
3. Lovable
Lovable is one of the fastest-growing darlings of the AI era. Currently at a projected ARR of $200M, experimental software developers (called vibe coders) have lifted the Swedish startup into the stratosphere.
The first time I tried Lovable, its homepage caught my eye. It’s simple, straightforward, and the homepage is basically one core hero message - “Build something lovable”
Lovable solves its users’ cold-start problem by animating the text field in its chat box. The text field shows the user all of the different things they could build with Lovable. Web apps, prototypes, internal tools, blogs, landing pages, etc.
This solves the cold start problem. In a world where users can build anything, the product should help guide their decision-making. Of course, users cannot test out Lovable without signing up - that would be a massive cost disaster.
The homepage goes further by showcasing web apps and websites from the Lovable community. These community projects serve as inspiration for new and existing users.
I would take this a step further by allowing users to duplicate or ‘remix’ some of those apps and websites.
4. Replit
Also spurred by the vibe coding revolution, Replit allows its users to build websites and apps. The company is projecting $1 billion by the end of 2026 and is seen as a more capable tool for building applications.
Replit’s homepage has a strong core message: Turn your ideas into apps. As a person who has all kinds of ideas all the time, this hit home for me. The page then goes further and shows an example prompt, immediately getting the user into the product. If the prospective user clicks the “show suggestions” button, the text changes, revealing more possibilities. These ideas are pretty cool!
The rest of the landing page shows case studies from enterprises that use Replit and showcases the product’s security features. Replit does not show what its community is building - this seems like a missed opportunity.
As with Lovable, a user cannot ‘test’ Replit without signing up.
If you’d like to try Replit, please use my unique link here. If you follow me on LinkedIn, you’d know that I have been building a few cool things with Replit’s agent.
5. Google’s Gemini
Currently the second-most-popular chatbot after OpenAI’s ChatGPT, Gemini has come a long way from its Bard days.
Its landing page is simple and straight to the point. Showing the user a text field to interact with the chatbot, while also showcasing what Gemini can do. Four clickable prompt suggestions can get the user started without requiring them to sign up.
This page enables product usage almost immediately and guides a first-time user from 0 to the aha moment.
The rest of the page does some heavy lifting. The user can change the page’s theme from a settings icon while clicking menu objects that showcase Gemini’s capabilities. In the chat box, Gemini restricts file attachments and model changes to signed-in users.
6. Figma Make
Figma’s Make is a Replit and Lovable competitor within Figma’s product suite. You can read a full case study of Figma here.
The product promises to “Make your ideas real with Figma Make”. The landing page places a chat box in the hero section and asks, “What do you want to make?” If the user is unsure, three prompt options are available.
The middle section of the page showcases web apps from the Figma community, while the rest of the page touts Figma’s features. The company is known for being community-driven, so it’s great to see what others are building.
There’s a vast difference between Figma and Lovable’s community showcase, though. Figma users are building very unique web applications - like a flower catcher game and a pattern generator while Lovable’s users build a lot of landing pages and websites.
7. Eleven Labs
I came across Eleven Labs while researching to build PLG motions for an AI project. This team absolutely gets Product-Led Growth. The company has products for creators, enterprises and developers. It showcases all of its capabilities within its hero message.
Prospects can test nearly all of its features. Text-to-speech, voice changes, music generation, speech-to-text, and audio dubbing. In just one section of the landing page, ElevenLabs uses PLG to its advantage, showcasing most of its features within constraints while also allowing users to generate speech, text, and music.
The rest of the landing page showcases social proof, use cases, and technical capabilities.
Conclusion - Showing Features > Telling
AI is not a strong enough differentiator for most software companies. Like software tools before them, showing is greater than telling. Enabling users to see how others use each tool or to interact with the tool itself is a far more effective way to drive adoption than simply communicating “AI” on the landing page.
Lovable and Figma get this right with community-centred landing pages. ChatGPT and Gemini allow users to test the basic features of the product without signing up. Replit showcases its capabilities via the suggestions button. Eleven Labs blows all of this out of the park by opening its product features to prospective users, showing the depth and breadth of its AI capabilities. Canva showcases its features appropriately via GIFs and images, but leaves much to be desired.
Overall, these platforms demonstrate practical approaches to driving product-led growth through new product features. Each page provides clarity on messaging and positioning to its intended audience, and I’m excited to see how AI features impact PLG.
Thank you for reading this Growth Case Study. If you loved this, please hit the share buttons and share with a friend or colleague. Please subscribe to receive the next case study.
Build smarter, not harder: meet Lindy
Tired of AI that just talks? Lindy actually executes.
Describe your task in plain English, and Lindy handles it—from building booking platforms to managing leads and sending team updates.
AI employees that work 24/7:
Sales automation
Customer support
Operations management
Focus on what matters. Let Lindy handle the rest.




Reply