1. The input is messy
A photo, short note, voice fragment, save, click, QR source, room, and timestamp arrive together.
Troviii demo: one Art Basel scan
A visitor sees something. Troviii turns it into a map.
The demo should be this simple: scan the room QR, capture the thing that moved you, let the AI read the signal, and leave with useful next steps. The host gets aggregate proof of what the room created.
Before
A visitor loves a work, takes a photo, forgets the booth, and never follows up.
Moment
They scan the room QR, photograph the work, and say: "I want the quieter version of this."
After
Troviii returns context, saves the intent, routes them to two relevant stops, and gives the host aggregate proof of interest.
Visitor entry
Art Basel
Desire Map
QR opens a room-specific capture surface. No app lecture first.
Camera capture
livevisitor photographs the work that stopped them
Magic moment
The AI reads the image and note, then Troviii turns it into: context, taste language, suggested stops, and a follow-up signal.
“You seem drawn to restrained craft and material memory. Save this work, visit the nearby textile installation, and ask the gallery for provenance.”
How it works
The demo arc: one messy human signal becomes useful follow-up.
A good demo should not tour every feature. It should make one transformation obvious enough that someone can retell it afterward.
2. The AI reads it
The AI extracts likely object, artist/place context, material, mood, taste language, and next-action intent.
3. Troviii structures it
The scan becomes a typed event signal: entity, context, intent, source, confidence, and attribution path.
4. The host learns
Repeated signals form clusters: what people noticed, trusted, wanted, skipped, asked, and followed up on.
Capture surface
The visitor should never feel like they are “using event software.”
They are doing the natural thing they already do at fairs: noticing, photographing, asking, saving, and comparing. Troviii turns that behavior into structured memory and useful next steps.
Scan room QR
No install. The guest lands on the specific booth, salon, dinner, or public-program surface.
Capture what moved them
Camera, voice, or note: a work, object, material, label, activation, room, or person.
Say the why
"I want something quieter," "send me context," "who else should I meet?"
Leave with a map
Saved memory, next stops, follow-up prompt, and optional share/compare card.
Live signal board
What the host sees
Input
event signalPhoto of ceramic vessel + note: "quieter version"
Room QR: salon B12
AI read
event signalceramic, Japanese craft reference, restraint, collector inquiry
confidence 0.87
Visitor payoff
event signalsaved to Basel map + two nearby works suggested
follow-up prompt ready
Host payoff
event signal6 visitors overlap on craft, restraint, finance, institution route
suggested small patron route
Technical object model
This is the infrastructure question: not a black-box moodboard, but a pipeline that converts raw multimodal inputs into typed event data.
Signal
raw input: QR, image, voice, text, click, save, visit, buy, share
Entity
work, object, artist, gallery, place, person, partner, public program
Context
event id, room, location, source person, time, role, consent state
Intent
save, inquire, meet, visit, buy, recommend, discuss, follow up
Attribution
who/what sparked the action and what downstream action followed
Privacy posture
The useful buyer output is aggregate and role-aware. Visitors can save privately; partners get trends and consented follow-up paths, not a creepy individual dossier.
Private saves by default
Aggregate partner reporting
Source and consent stored with signal
Paid merchant access gated by entitlement
Onboarding
Collectors, artists, designers, and institutions each get a simple intake.
Collectors and patrons
Input
Private collecting lens, works saved, advisor notes, salons attended.
Output
Personal Basel memory map, warm intros, post-fair revisit list.
Artists and galleries
Input
Artwork context, studio notes, provenance, available works, inquiry path.
Output
Interest graph, repeated questions, patron follow-up, shareable work card.
Designers and architects
Input
Materials, objects, rooms, routes, showroom moments, collaborators.
Output
Design trail, saved references, inquiry intent, collaboration leads.
Institutions and city partners
Input
Public programs, museums, city routes, public art, partner activations.
Output
Civic memory layer: what people noticed beyond the fair floor.
Visitor output
A personal Basel map: works saved, rooms visited, people to follow up with, routes to revisit, and one shareable card if they want to compare with a companion.
Partner output
Aggregate room intelligence: attention clusters, trust paths, product/work intent, underused routes, suggested follow-ups, and what should happen after fair week.