Unique Research · Insight / Unique Friend Q&A · October 10, 2024 · Guest: CEO, Zion.AI / Hanzai Technology (函子科技)
Product: Let users build their own web apps without learning to code or heavy professional training. No direct domestic peer; for a non-technical founder who needs to validate an idea fast, Zion is positioned as the highest-efficiency choice.
GenAI on design / GTM / pricing: Design — short term: users make and commercialize their own AI products; long term: a tools copilot, bottom-up. GTM — AI can draft replies, but Zion uses it only for intent classification so humans answer; AI also monitors Reddit/Twitter and competitor (Bubble) forums. Pricing — added AI usage fees by token volume.
Tech lead: LLMs are a new software primitive — like databases or React. Good product design lets users exploit and commercialize that capability; focus is functional orthogonality so DB/frontend/etc. interop with models (agents). Bottom-layer tech will not observably jump in 3–5 years (CEO predicted the post-LLM path in May 2023; still roughly on track).
Ecosystem: Positioned as a developer tool — anything that raises efficiency or lowers difficulty should be built; every useful integration should be supported; prefer standards, especially existing code standards.
Challenges / markets: Collapse in China startup formation count + missing overseas content. Domestically the only widely felt “AI that works” may still be Feishu meeting notes; overseas, calling a Lexus dealership, the IVR/human booking slot is already AI-replaced.
2024 watch: Less interest in content generation (weakly tied to the business model — “just an API,” linear change). Model miniaturization; more common multimodality; OpenAI’s Realtime API enabling new interaction modes.
— END —