Exclusive Interview with Guo Lie: After Creating Hits Like Lianmeng, FaceU, and CapCut, He Explains for the First Time How to Build Products in the AI Era 专访郭列：做出脸萌、FaceU、剪映等爆款后，他第一次阐释如何在AI时代做产品

TL;DR

• Flova.ai introduces a dual-audience product design philosophy, simultaneously optimizing interfaces for human creativity and Agent structural requirements.
• The platform utilizes a tri-view panel (Dialogue, Storyboard, Preview) to bridge the contradiction between unstructured human intent and structured Agent context.
• Context management is prioritized over simple memory, enabling Agents to retain semantic continuity across hundreds of dialogue turns and multiple episodes.
• Founder Guo Le leverages his "Jianying lineage" experience to apply rigorous A/B testing and data-driven growth strategies to AI Agent products.
• Flova distinguishes itself from model showcases and manual workflow integrations by offering a true Agent that inherits creative decisions, lowering barriers for non-professional creators.

Why It Matters

This article provides a critical framework for AI product designers, highlighting the unique challenge of creating tools that serve both human users and autonomous Agents. It demonstrates how successful AI ventures are moving beyond simple API wrappers to build sophisticated, state-aware environments that facilitate seamless human-Agent collaboration. For investors and practitioners, it underscores the importance of strategic capital allocation and the shift toward data-driven decision-making in the early stages of AI startups.

Technical Details

• Dual-Interface Architecture: The product features a unified workspace with three distinct views: a "Dialogue" area for natural language instructions (Context input), a "Storyboard" area for visualizing Agent execution, and a "Preview" area for real-time visual feedback. This allows humans to interrupt and adjust workflows dynamically.
• Advanced Context Management: Unlike basic memory retention, the system focuses on managing complex project contexts. It tracks semantic consistency across long conversations (hundreds of turns) and multi-episode projects, ensuring style and narrative coherence without requiring users to re-prompt.
• Agent-Centric Workflow: Flova operates as a third-tier AI video tool, distinct from C-end model showcases or manual multi-model integrations. It automates scriptwriting, character design, storyboarding, and editing through encapsulated Agents, reducing the need for user expertise in prompt engineering or workflow assembly.
• Data-Driven Product Development: The team employs systematic A/B testing and internal observation tools based on shared Contexts between product, R&D, and operations. This replaces intuitive guesswork with scientific validation of feature efficacy and user behavior.

Industry Insight

• Redefining Product Design for AI: Traditional UI/UX principles are insufficient for Agent-based products. Companies must design "containers" that provide structure for Agents while allowing flexibility for humans. The success of Flova suggests that the next wave of AI applications will be defined by how well they manage this bidirectional information flow.
• Strategic Capital Efficiency: Guo Le’s decision to raise only $80 million despite high demand challenges the industry norm of excessive fundraising. This indicates a maturing market where sustainable business loops and product-market fit are valued over cash burn, suggesting a more disciplined approach to startup valuation and growth.
• Shift from Tool to Partner: The evolution from manual AI tools (where humans orchestrate models) to true Agents (where humans set goals and Agents execute) represents a fundamental shift in user engagement. Products that can effectively lower the barrier to entry for complex tasks like video production will capture the mass market, not just professional creators.