M Hudson Design × aiChemist · direction decisions

Six calls. Five minutes.
Every one has a recommended default.

Context for each question is in the briefing. Pick an option (the recommended one is marked), add a note if you want, and hit send at the bottom.

0 of 6 answered
Decision 1 · Brand model

Does the trained brand style read as "M Hudson"?

Your visual signature — materials, planting, lighting — is now trained into the model itself (samples at the show-and-tell). v2 retrains on your real photography for sharper realism, minus the watermark artifact.

Decision 2 · Exhibit styling

Client project pages: dark gallery, or light like the website?

Public pages already match your site's light botanical brand. The client exhibits can either go dark-sage (gallery drama, on-palette) or the same light treatment.

Decision 3 · Visibility

Keep the client project pages public, or gate them again?

Connors / B. Davis / Wiener exhibit pages are currently open to anyone with the link (no client names appear beyond project labels). Re-gating takes minutes.

Decision 4 · Cost posture

Keep the current cost posture?

69% of your archive's finished variations cost $0 on the in-house GPU. The paid model runs only where measurement proved it necessary — about 13–24¢ per image at those specific points.

Decision 5 · Show-and-tell

Run the proposed 7/7 sequence?

Proposed: your plans rendered → any drawing from anywhere → your 509-still archive restyled and browsable → walk inside a design → the pick-a-mood interaction your clients will drive.

Decision 6 · Real-property walk

Green-light the first capture-grade walkthrough?

Today's 360s are AI teasers. The real product is a scanned, navigable model of an actual property — Carlos's site capture feeds it. Picking the first property starts that track.