How a ‘vibe working’ approach at Genspark tripled ARR growth and supported a barrage of new products and features in just weeks

In a landscape where traditional product launches can be burdensome and bogged down by bureaucracy, Genspark is redefining the game. This AI workspace company has embraced a novel strategy known as 'vibe working,' allowing its lean team to operate at what they term 'gen speed.' This approach has enabled Genspark to roll out new products and features almost weekly, significantly boosting their annual recurring revenue (ARR). According to co-founder and CTO Kaihua (Kay) Zhu, this AI-native methodology empowers every team member to take on managerial roles, supported by a team of AI agents. These agents help deliver features end-to-end, streamlining the development process. Launched in June 2024, Genspark initially focused on AI search but quickly shifted gears to its flagship product, Super Agent. With 5 million users, Super Agent goes beyond traditional static search by dynamically selecting the best tools and sub-agents, optimizing results in real time. Set to launch on April 2, Super Agent leverages Anthropic’s Claude AI technology, which Zhu claims can compress an afternoon of office work into just five minutes. The tool is remarkably versatile, capable of performing tasks ranging from making calls and fact-checking to drafting documents and conducting extensive research. Zhu, a veteran with over two decades in search technology, asserts that while it still functions as a search tool, it is far more advanced than its predecessors. Genspark has quickly expanded its feature set, launching a series of competitive challenges to highlight its capabilities. Following a comparative analysis with OpenAI's ChatGPT agent, Genspark initiated a '1 Million Dollar Side-by-side AI Showdown,' inviting users to identify instances where other platforms outperformed Super Agent. The first round revealed 429 cases where ChatGPT triumphed, rewarding participants for their findings. In a subsequent round, Genspark increased the stakes, offering $200 for each successful challenge against any AI tool. The company attributes its rapid growth to its compact yet powerful team of 20, driven by an engineering philosophy that emphasizes 'less control, more tools.' Zhu explained that over 80% of Genspark's code is generated by AI, although they maintain a stringent code review process to ensure quality. This allows the team to function efficiently, akin to superheroes in a collaborative effort, while also preparing to scale as needed. Zhu advocates for a restructured approach to modern enterprise teams, having experienced the pitfalls of larger, more hierarchical organizations. Genspark promotes transparency and high productivity, with every team member focused on delivering shippable products. The team embodies their product philosophy, engaging with the tools they create to continually refine the user experience. The evolution of Genspark's technology reflects a broader trend in AI, where rigid workflows are giving way to more flexible, adaptive systems. Super Agent employs a mixture-of-agents (MoE) framework, utilizing multiple large language models to break down tasks and ensure accuracy through cross-verification. With access to over 80 specialized tools and diverse datasets, Genspark aims to enhance accuracy while minimizing errors. Zhu acknowledges that while Genspark is not striving to create cutting-edge systems like other industry leaders, their focus remains on efficiency—reducing costs and latency for simpler tasks. Their vision is to democratize the coding experience, making it accessible even for those unfamiliar with programming. Genspark is committed to fostering an environment where everyone can explore and experiment, embodying the spirit of 'vibe coding.'