Working product vision
Kyshik
This deck describes a direction for a pilot. It does not present unverified model accuracy, customer commitments, or revenue projections as current results.
01 · The idea
Verified records for animals seen in the field.
A shared record can connect sightings and care over time.
02 · The problem
Animal histories are fragmented.
Photos, locations, and care notes often remain in phones, chats, and spreadsheets.
03 · The approach
Use a phone, a photo, and a current location.
A similarity model can suggest a possible existing record; staff review important matches.
04 · The record
One profile, many dated sightings.
Photos and observations build a longitudinal field record.
05 · The workflow
Capture, review, coordinate, follow up.
Field teams and coordinators use a map to plan humane follow-up.
06 · Privacy
Exact locations stay in the staff workflow.
Public views use aggregate counts and adoption listings without private map coordinates.
07 · Care records
Record each action as an event.
Sterilization, vaccination, treatment, and adoption need dated evidence before public claims.
08 · Public value
Clear counts with clear definitions.
Sightings, profiles, and verified care events are different measures.
09 · Model quality
Measure before relying on a score.
Local, labeled dog and cat photos are needed to validate matching thresholds.
10 · Field pilot
Learn from a small, supervised rollout.
Test permissions, camera, GPS, offline sync, duplicate handling, and staff review.
11 · Partnerships
Coordinate with shelters, clinics, and city teams.
A pilot should define who can access records and who confirms care events.
12 · Next steps
Build a trusted system of record.
The current product is an early prototype; coverage, accuracy, and operating costs need pilot evidence.