Lenny's Podcast: The new AI growth playbook for 2026 — Elena Verna
What the video covers
Published December 18, 2025, on Lenny Rachitsky’s YouTube channel, this conversation features Elena Verna, Head of Growth at Lovable — an AI-powered app builder that reached $10 million ARR in sixty days and $200 million ARR in under a year with around 100 employees. Verna spent years as a growth consultant working with SurveyMonkey, MongoDB, and Amplitude, which gives her an unusual vantage point: she has seen traditional growth work at scale and now leads growth at one of the fastest-growing AI companies on record. The conversation focuses on what has stopped working in AI product growth and what is replacing it.
Who it’s for
Product leaders, founders, and growth practitioners at companies building AI products, particularly those who have tried applying conventional SaaS growth frameworks and found them producing weak results. Also relevant for PMs in larger companies evaluating whether standard growth and retention playbooks apply to AI features they are shipping.
Key takeaways
1. Most of the SaaS growth playbook does not transfer. Verna estimates that 60 to 70 percent of traditional growth tactics no longer apply to AI products. The root cause is structural: in a SaaS product, the feature set is relatively stable and growth work optimizes the path to value. In AI products, the capability changes substantially every few months, which means the path to value changes with it. Growth tactics calibrated to one capability state become obsolete when the model improves.
2. Product-market fit is a moving target. Verna describes AI companies needing to “re-find product-market fit every three months.” This is not pessimistic — it reflects that product capabilities genuinely expand fast enough that what users wanted six months ago may now be achievable differently or may have been superseded entirely. The consequence is that growth and product teams must stay in close contact and continuously retest their assumptions about what drives retention.
3. Activation belongs to product, not growth. Traditional growth teams often own the onboarding funnel and activation metrics. Verna argues this no longer makes sense for AI products, where the user’s ability to experience value depends on product decisions — model selection, interface design, context management — that growth cannot influence. At Lovable, activation was moved into product ownership, and growth focus shifted toward continuous feature shipping and experimentation.
4. Free outperforms paid acquisition. Rather than running paid advertising, Lovable’s most effective growth lever was giving the product away for free. Verna explains this in terms of the discovery surface: when a user builds something with Lovable and shares it, the output carries a provenance link. This creates organic distribution that paid channels cannot produce at equivalent cost or authenticity. The “minimum lovable product” framing — a standard of actual user delight rather than minimum viability — is part of what made the free distribution work.
5. Speed of iteration is itself the strategy. Cursor, another AI-native company cited at $300 million ARR, follows a similar pattern: the product improved fast enough that word-of-mouth kept pace with growth. Verna frames this as a structural shift: when capability development is fast enough, it becomes the primary growth mechanism rather than a background condition. This changes the resource allocation logic — investment in product iteration has a higher marginal return than equivalent investment in traditional growth infrastructure.
Worth watching if
You are responsible for growth or retention at an AI product and are finding that your usual frameworks are not producing expected results. The Lovable and Cursor data points give the argument concrete reference points rather than leaving it as theory.