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

Host: Teresa Torres
Guests: Vlad (CEO & Co-founder), Vova (CTO & Co-founder), Vlad (Founding Growth)
Date: April 2, 2026

Episode Overview

This episode explores the journey of Banani, an AI-powered, canvas-first product design tool aimed at making high-quality design accessible and faster for everyone. The founders share their insights on why design is a massive differentiator, the technical and UX challenges they faced, and how an AI design agent can collaborate with humans without eliminating the need for human creativity. The conversation covers Banani’s origin story, prototyping, agent workflows, maintaining design quality, and their unique product philosophy in the crowded AI design space.

Key Discussion Points & Insights

1. Banani’s Mission and Product Vision

• Purpose: Banani aims to automate and expedite much of the design process by providing an AI-powered design companion that augments human creativity—especially valuable for teams or founders who lack access to experienced designers.
• Unique Approach: The product is "canvas-first," focusing on giving users both the freedom of manual design and the generative power of AI, as opposed to code-first or chat-first tools.

Quote:
“We try to build [an] AI product designer because... it is so hard to design user flows, come up with best UX solution... Our goal is to build a fully autonomous agent that covers everything related to design inside your company, something that can be your creative companion.”
— Vlad, CEO (01:26)

2. Founder Motivation and Raising the Floor on Design

• Motivation: Both Vlad and Vova draw from their backgrounds in design, product management, and engineering. Their personal pain points, especially being stretched thin as the sole designer in fast-moving startups, motivated them to pursue this idea.
• Philosophy: Rather than replacing designers, Banani raises the baseline quality of design (“raises the floor”) so designers focus on high-level problems instead of repetitive tasks.

Quote:
“I think there’s kind of irreplaceable parts on the human side, which is empathy... But sometimes you just need things that can help you go through creating thousands of screens in a day. Raising the floor is also a big part of that.”
— Vlad, CEO (04:17)

3. Spotting the Problem and Product Genesis

[06:11]

• Inflection Point: The arrival of accessible generative models (GPT, LLMs) made it feasible to experiment with automating design, moving beyond code gen to more creative domains.
• Initial Solution: They began with proof-of-concept Figma plugins to validate demand and experimentation before investing in a standalone product.

Quote:
“We made proof of concepts to just see if it's feasible to do design generation…our first kind of working product was plugin for Figma...after validating that the tech is feasible...it was the first signal that we can try [to] build a full on product.”
— Vlad, CEO (11:05)

4. Differentiating in a Crowded Space

[16:24–19:46]

• Opinionated Design: Banani maintains a design-focused, opinionated viewpoint (e.g., click-to-generate-next-frame) rather than following developer-leaning paradigms.
• True Canvas-First: Unlike code-first tools that generate applications or chat-based AI that sits outside primary workflows, Banani integrates deeply with the designer's UX mindset and workflow.

Quote:
“Nobody solved the problem of generating actually good, tasteful, beautiful design... And the good thing here is that the market is extremely big... But for example, magic patterns and other stuff...they pretty much in a lot of ways [are] developer focused...”
— Vlad, CEO (16:24)

5. Product as it Stands: Experience, Autopilot, and Collaboration

[20:27–26:52]

• UI/UX: Users interact via a canvas, blending direct manipulation (drag, select, edit) with agent-powered generation.
• Agent as Copilot: The AI is positioned as an autopilot that can "take the wheel" when the designer wants, but never takes full control, preserving user agency.
• Customization for Design Stages: The AI’s level of autonomy adapts to the user’s phase—creative exploration or repetitive production.

Quote:
“What if the AI is not here to replace... the product designer,...but rather to be autopilot where the designer is still in the driver's seat, but when...needs to switch to autopilot, they can do it.”
— Vlad (Growth) (24:08)

6. Technical Deep Dive: Architecture, Agents, and Tooling

[28:04–56:53]

• Behind the Scenes: The system splits user requests into sub-tasks, manages design states/history, and supports parallel edits across screens.
• Key Differentiators:
  • Not “code-first” but “mockup-first,” prioritizing the look/feel rather than immediate code execution.
  • Maintains robust session and version histories, enabling reversible workflows.
  • Ongoing experimentation to refine context management, prompt engineering, and agent tools.
• Quality Strategy: Heavy use of context engineering, tool orchestration, and postprocessing for aesthetics; willingness to accept minor model limitations short-term to focus on more enduring product values.

Quotes:
— "Not everything that happens under the hood and not everything that we share with our agent as a history of the session is actually like a tool call... But sometimes we mention it to our agent as a tool call because...the data is consistent, it is easier to understand."
— Vova, CTO (49:27)

— "50, 50 work between, once again, how you orchestrate stuff and what things you provide on the app layer. And yeah, like combination of both of those can make or break it for your products, for your company."
— Vlad, CEO (51:36)

7. Handling Ambiguity, Context, and The Human-AI Interface

[36:11–43:19]

• Communication Gaps: Users and agents “speak different languages,” especially challenging in design where visuals are hard to describe in text.
• UX Innovations:
  • Select elements or highlight areas on the canvas to clarify intent.
  • Agent can proactively ask clarifying questions ("Do you mean that element?").
  • Predefined, quick actions for common tweaks.
• History & Context Windows: Each screen maintains its own history/context, aiding complex multi-screen projects.

Quote:
“We have this huge gap between how people communicate with LLMs and agents and that how agents can comprehend all of that stuff. And, you know, it's like it's okay-ish when you write in code, but it's not okay when you're building design..."
— Vova, CTO (33:48)

8. Evolving with Technology: Betting on the Future

• Strategic Focus: They consciously avoid over-investing in “solvable” problems like alignment bugs, banking on models improving, and instead direct resources at workflow unlocks that won't be “eaten” by model improvements.
• Long-Term Vision: Building for where they believe the industry is headed in 6–12 months ("pulling the future forward") and dreaming about transformative, local-first AI apps.

Quote:
"It’s really important to dream of the things that are not even possible. And I think this is how you should create your vision of the company. It's not always about what should and can we build today."
— Vova, CTO (62:09)

9. Memorable User Story

[68:01]

• Vova shares a personal tale of using Banani to finish a real work project, finding it superior to working directly with legacy tools:
  “The real thrill for me is the moment when I understand that I can use the product I built... I've imported design from Figma into Banani... And this was amazing experience for me.”

Notable Quotes & Moments with Timestamps

• On the core vision:
  "[Banani is] your creative companion. If you're a designer yourself...our goal is to build a fully autonomous agent..." — Vlad, CEO (01:26)

• On raising the floor of design:
  "Nobody likes to use bad products...we just want for the world to have more access to get to the great and quality user experience, user interfaces." — Vlad, CEO (05:19)

• On the agent as autopilot:
  "What if the AI is not here to replace...the product designer...but rather to be autopilot where the designer is still in the driver's seat..." — Vlad, Growth (24:08)

• On experimenting and evolving:
  “There’s no easy way outside of just trial and error... Even right now when each new model comes out, you need to get a taste of what it can do...” — Vlad, CEO (28:04)

• On dreaming and building for the future:
  “It's really important to dream of the things that are not even possible. And I think this is how you should create your vision of the company." — Vova, CTO (62:09)

Timestamps for Key Segments

• Origin Story & Team Introductions: 00:09–01:21
• What is Banani: 01:26–04:17
• The Problem Space & Why AI Now: 06:11–08:50
• Figma Plugin and Early Prototyping: 11:05–13:12
• Distribution Strategies & Growth: 14:17–15:33
• Differentiation & Market Dynamics: 16:24–19:46
• Current Product & Canvas-First UX: 20:27–23:48
• Autopilot Philosophy & User Control: 24:09–27:14
• Maintaining Quality & Craft: 28:04–33:02
• Bridging the Human–AI Communication Gap: 36:11–43:19
• Behind the Scenes: Agents, Tools, Context Management: 43:26–56:53
• Balancing Short-Term Gaps vs. Long-Term Vision: 60:22–66:32
• Founder Anecdote on ‘Dogfooding’: 68:01–69:27

Final Thoughts and Ways to Learn More

• Start using Banani at banani.co (67:34)
• Free tier available for experimentation by new users.
• Continuous enhancement of the agent and focus on solving creative collaboration challenges.

“The real thrill for me is the moment when I understand that I can use the product I built.”
— Vova, CTO (68:01)