AI Software's New Playbook

Why instant wow, viral marketing, and simplicity define today’s software winners

I’ve spent the last three years building software and investing in AI-first startups like Lovable, Fyxer, Parahelp, Glif, Superscale, Harmony, Fixkey, and many more. 

In recent months, I've witnessed a seismic shift in the software landscape. Here are some random thoughts and observations about the current state of software and AI:

1) Simple products with a quick time-to-wow win

AI is reshaping not just how we build software but also what kind of software we build. Users are now flooded with choices, and the products that break through have one thing in common: a super-short "time to wow." The fastest-growing tools reveal their value instantly.

Take Perpexity which gives you a flawlessly written, sourced response in seconds. Or Lovable, one prompt and you start building a full web product with a database with. Or Fyxer, which clears your Inbox and drafts responses to unanswered emails within moments after you connected your Gmail.

Chat interfaces have set a new standard—users expect immediate results. The patience for figuring things out has vanished; if it takes a few minutes to set up or steep learning curves, and your tool risks swift abandonment.

2) Companies grow faster than ever—but will they stick around?

We’re seeing incredible revenue numbers, like startups hitting $20M ARR within months. When we built Blinkist, 3x revenue year-over-year was considered very good, and VCs got excited. Now, some VCs won't even look at startups if they are under $10M ARR.

I really like that we shift the definition of success to actual revenue instead of vanity metrics like MAU. But I'm also a bit cautious about the sustainability of these numbers.

Today’s AI landscape feels a bit like the early days of the App Store when “there was an app for that”. Back then, users were still downloading several apps per month, but over time, only a few stuck around. 

I think we’re experiencing something similar with AI software, just with one main difference: people are willing to pay for AI software. This is very different from the early App Store days when people complained even about having to spend $0,99 on an in-app purchase.

My big question is the stickiness of the revenue. With new AI tools emerging daily, will users remain loyal, or will they quickly move on and chase the next shiny thing?

I'm not saying this as a bubble; many companies are genuinely building robust businesses with solid retention. Yet learnings from the previous cycle suggest that many startups’ growth will slow down eventually. I sincerely hope this time is different, and everyone emerges a winner.

3) Distribution is changing

This shift may prove even more significant than AI itself.

Today's fastest-growing software startups aren't relying on traditional channels—they’re leveraging social media, influencers, and viral marketing. Awareness is now king. With many tech companies and startups making big release announcements nearly weekly, cutting through the noise is critical. Traditional channels—PR, paid marketing, SEO—still work, but often lack the speed and scale needed today.

A new breed of founders understands this intuitively. They’ve grown up online, built their brands publicly, and cultivated communities organically. Elon Musk famously pioneered this approach (claiming Tesla never spent a penny on marketing), and Sam Altman openly jokes he's OpenAI’s marketing department. Now, founders like Anton from Lovable, Carl from Nothing, or Roy from Cluey take this to the next level. They inherently grasp virality and now how to create awareness (and controversy).  They're designing product features specifically optimized to perform well on platforms like TikTok and Instagram.

“If your head of marketing doesn’t have 100k followers on social media, they're not gonna make it.” – Roy from Cluey.

It’s a new world, and mastering this new form of marketing is no longer optional if you want rapid growth.

4) The software stack is changing.

I recently talked to a lot of people about their software stack, and I noticed that many people are moving back toward simple setups. Instead of endlessly searching for off-the-shelf software solutions, people are keeping their main tool stack simple and then vibe-coding small tools to solve their specific, individual problems.

In the past, users would spend time researching software tools to solve their problem and then buy different products for it. Now they are solving their problems themselves and often abandon CRM or workflow software tools.

They’re stitching together data from various sources to create custom web views or setting up personalized chatbots with their data. 

It’s one of the first substantial use cases of "vibe coding" I've observed in the wild. Users aren’t just buying software anymore—they’re actively shaping it themselves.

This shift isn’t minor. It fundamentally changes how we think about software buying and might redefine software creation for years to come.

Photo by itsiken on Unsplash