Lumexa is an agent-native automation platform: watch a task once on screen, and it executes forever, no integrations, no code. The vision slide says it plainly - the future of work has no APIs.

The problem with a vision that large is that it doesn't answer the second question investors ask. Lumexa's revenue wasn't coming from the future. It was coming from TKK Co. Ltd., converting 2D CAD drawings into 3D models, at $200K average annual contract.

So the work split into two decks rather than one compromise. The seed deck carries the horizon: SLTM as memory infrastructure, a three-phase roadmap to 2028, TAM/SAM/SOM, use of funds against a $20M raise. The investor deck stays on the floor: real savings per task, competitive moat in visual reasoning, bottom-up market math, path to $100M ARR.

Both run on one system a single accent green against deep forest, pill-shaped data containers that make numbers legible at a glance, and a mark that scales from favicon to full-bleed. Traction stays loud throughout: $1.3M contracted, 1,000+ waitlist, Sompo Japan.

Deck design isn't decoration. It's deciding which argument each room needs.