Intro
Tan Luc’s AI workflow product work centers on a simple idea: the model call is not the whole product.
For tools like CinematicAI, the useful surface is the workflow around generation. Users need to see inputs, steps, outputs, failures, and reusable patterns.
Product shape
An AI workflow product should make structure visible. Prompt boxes can be fast, but they often hide the process people need to inspect later.
When the product exposes flow, state, and relationships, users can debug work instead of guessing what happened.
Frontend role
Frontend systems matter because they carry the user’s mental model. Layout, state, node boundaries, loading states, and failure states all decide whether the tool feels understandable.
The best AI workflow UI does not decorate the model. It makes the work legible.
Direction
For luc.works, AI workflow products are a long-term theme: practical tools for creating, inspecting, repeating, and improving AI-assisted work.