Building Banani: How a Canvas-First AI Designer Is Raising the Floor on Product Design

Banani was founded by Vova (CTO) and Vlad (Growth Lead) after recognizing gaps in design resources for startups and individuals. Their prior work on consumer apps underscored the importance of design as a competitive differentiator, leading them to create an AI-driven platform aimed at automating design workflows. The company seeks to empower non-designers by generating scalable, high-quality designs weekly, with a focus on improving user experiences through AI-assisted creativity. While acknowledging AIs current limitations in empathy and emotional nuance, Banani emphasizes its ability to handle routine tasks, freeing designers to focus on complex challenges. The platform also prioritizes aesthetic quality to avoid generic, unappealing outputs often associated with AI tools.

Bananis product development leveraged early experiments with LLMs, such as GPT 3.5, to validate design generation feasibility. A Figma plugin served as the first practical test, enabling rapid prototyping and iterative improvements. The platform distinguishes itself by offering a canvas-based interface, allowing users to edit designs directly and generate full flows via text prompts, rather than relying on code-centric tools. It addresses pain points in existing design tools by integrating AI with design systems, supporting parallel editing, and balancing automation with human oversight. However, challenges persist, including aligning AI outputs with design principles, refining context for better results, and bridging communication gaps between human intent and agent interpretation. Banani aims to streamline workflows for prototyping, reduce manual labor, and democratize access to high-quality design while maintaining a focus on aesthetics and usability.

Looking forward, Banani seeks to enhance AI-human collaboration by making agents more proactive in interpreting user needs and refining workflows. Future goals include enabling full autopilot for low-uncertainty tasks, such as generating prototyping flows from product requirements, and integrating AI agents into broader creative processes. The team also plans to expand interaction modes beyond text-based commands, experimenting with graphical selections and hybrid interfaces. Despite advancements in LLMs, challenges remain in achieving high-fidelity outputs and ensuring AI complements rather than replaces designer expertise. Bananis vision balances automation with creative control, aiming to redefine design workflows through AI while addressing the evolving demands of the design and prototyping markets.