Only 30% of growth experiences are applicable in the internet era! Lovable: AI product cycles iterate every three months with LLM

AI Coding Platform Lovable is rewriting growth records at an astonishing speed. This startup, founded just a year ago, has achieved $200 million in annual recurring revenue despite having fewer than 100 employees, and recently raised $330 million in Series B funding with a valuation of $6.6 billion. Investors include NVIDIA. Elena Verna, Head of Growth at Lovable, revealed how the company is abandoning traditional marketing doctrines and redefining growth strategies in the AI era during an exclusive interview on the well-known podcast “Lenny’s Podcast.”

The AI Era Requires Abandoning Old Scripts: From Optimization to Innovation

Elena Verna pointed out that companies like Lovable at the forefront of the AI wave have only 30% to 40% of their traditional growth experience from the past 15 to 20 years still applicable.

She said, “In the past, I spent 95% of my time on optimization and 5% on innovation; but at Lovable, I spend 95% on innovation and only 5% on optimization.” Verna emphasized that in the highly competitive Vibe Coding field, relying on micro-tuning conversion rates or simply changing button colors is no longer enough to stand out. True growth comes from continuously launching new features (such as integrations with Shopify or voice modes), which serve as the core growth engine.

The Harsh Reality: Reclaim PMF Every Three Months

Verna presented a sobering perspective for all AI entrepreneurs: Product Market Fit (PMF) is no longer a one-time milestone.

Because AI models (like OpenAI, Anthropic) improve dramatically every few months, combined with rapidly changing consumer expectations, AI companies must re-establish PMF every three months. This means companies cannot rely solely on sales teams for expansion; product teams must always be in a state of high alert to keep up with rapid technological and market iterations.

Long-term Roadmaps Are Dead, Only Look at 3 Months

The old concept was to develop product roadmaps for 1 year or even 3 years. But now, 3 months equals 1 year in the AI era. Due to the rapid iteration of foundational models (such as GPT-4, Claude 3.5), any planning beyond 3 months is futile. After 3 months, the technological boundaries will shift, and user expectations will change. Companies must be highly agile, recalibrating their direction every quarter.

User Experience: Aha Moments Are Dead, It Must Be Wow Moments

The old idea was to guide users through a series of operations until they discover the product’s value (Aha Moment). But in reality, users lack patience. AI products must make users feel “cognitive explosion” at the very first prompt. This is no longer about understanding value through an “Aha,” but about being struck by a “Wow Moment.” If the first interaction isn’t stunning enough, users will churn.

Abandon Sales Teams, Let the Product Sell Itself

When Lovable reached $200 million ARR, it had almost no traditional sales team (only a few people handling passive demand). Relying on exceptional product-led growth (PLG) and word-of-mouth, users become the best salespeople themselves. In the AI era, the product itself is the strongest salesperson.

Abandon Profit Margins, Now Is the Time for AI Land Grab

AI model token costs are indeed high, leading to lower gross margins. But now is not the time to make money; it’s the time to aggressively capture market share (Land Grab). Consider the expensive AI computation costs as customer acquisition costs rather than operational costs. Saving money now will lead to obsolescence in the future.

Abandon Late Majority, Focus on Pioneers

AI development is too fast for most late majority users to keep up. Trying to please everyone results in mediocrity. Currently, the focus must be on serving the most enthusiastic and technically savvy pioneers, because only they can keep up with every three-month technological iteration and provide cutting-edge feedback.

There’s No Such Thing as Too Big to Fall, Only Too Slow to Survive

AI moat is extremely fragile. Even absolute leaders like OpenAI could see their market share collapse within weeks if models like Gemini or Claude outperform theirs (OpenAI recently lost share due to Gemini 3). There’s no such thing as too big to fall, only too slow to adapt.

This article: Only 30% of Growth Experience in the Internet Era Applies! Lovable: AI Product Cycles Iterate Every Three Months with LLM. Originally published in Chain News ABMedia.