SAAI PRODUCTS
The answer your space holds, SAAI.
SAAI is an AI built to read space.
From the cameras already installed it reads behavior rather than people, and proposes what to change with the grounds for it. Experience stays in the space, not in the payment record. A store, a showroom, a distribution center.
Uses existing cameras · no identity retained
Video is anonymized at input (SEAL), and the original is never stored.
More branches. Still three things to watch.
Where to act now. Why it changed. What to move next. Three products answer on one floor plan.
TODAY · YESTERDAY · TOMORROW
Pick a space and the floor plan, the routes, and the stage names switch with it.
Convenience store
Convenience-store flow
A simulation rebuilt from patterns observed in real operations. It reflects the space layout and movement conditions.
What this space asks
- 01
How many visitors did an out-of-stock shelf turn away
- 02
Does the entrance-front display carry people deeper into the aisles
- 03
Where do people stop when the checkout queue grows
The three panels use the convenience store as their example.
Which branch needs a hand right now?
- Fridge temp exceeded2 min agoE · Chilled aisle
- Stockout11 min agoD · Aisle 3
- Queue 3.2 minOne more staff member, 4 to 6 pmE · Chilled aisle
What it reads
- crowding
- queue
- stockout
- anomaly
What you getLive detection and alerts
Look at saai careWhy do branches end up different?
What the zones say
- A · Wall aisle
- Glanced at in passing
- B · Aisle 1
- A destination aisle people seek out
- C · Aisle 2
- Where planned purchases cluster
- D · Aisle 3
- Frequent traffic, short dwell
- E · Chilled aisle
- Highest traffic, most contested reading
One day in five stages
- 1,160Passing
- 382Entry33%
- 317Dwell
- 121Pick-up
- 65Purchase
Illustrative figures modeled on a convenience store.
What it reads
- pass-by
- capture
- path
- dwell
- gaze
- pick-up
What you getTime series, patterns, hypotheses
Look at saai insightWhere do we move what worked?
Run it back under the same conditions and compare. The person operating decides.
Compare both layouts under the same visit conditions. Pick the alternative and the map updates immediately.
What it reads
- drop-off
- return
- restocking
- staff conversation
What you getPriorities and proposed actions
Look at saai agentThree screens.
Today as an alert list · yesterday as a dashboard · tomorrow as a head-office table.
saai care
saai care · HQ217 stores nationwideLIVE37
Anomalies today
94%
Hygiene·temp met
14
Open alerts
4
Visit needed
Store ranking · incidents
Store 7-day Status Sinchon 13 At risk Anyang 6 Watch Gangnam Stn. 4 OK Bundang 3 OK Live alerts
- Fridge temp exceededSinchonAt risk
- Floor spill detectedSeomyeonWatch
- Exit blockedAnyangWatch
By type · top 4 stores · 7 days
Hygiene11Equipment7Theft·anomaly5Safety3OKWatchAt riskAnonymized by SEAL · no footage stored* Sample screen · illustrative data Today · sense. The store that needs a hand rises to the top.
saai insight
saai insight
TodayStore Analytics Dashboard
Visits peak at 2–4 PM
- Est. daily sales
- 0.0k
- Total visitors
- 0
- Avg. dwell time
- 0.0min
- Shelf conversion
- 0%
Beverage shelf dwell time is 2x the usual. New-product promotion effect confirmed.
app.saai.ai/insight/dashboard/gangnamsaai insightGNGangnam
Pro plan
Store Analytics Dashboard
Visits peak at 2–4 PM
VisitorsPrev weekAI forecastZone heatmap
LiveLowHighAI Insight Report
Beverage shelf dwell time is 2x the usual. New-product promotion effect confirmed.→ Display-expansion action card created
Store ranking
All 12 storesYesterday · analyze. Same period, same zone, side by side.
saai agent
saai agent · HQ217 stores nationwideLIVE1,145
Visits today (top 4 stores)
19.7%
Avg conversion
82%
Calls acted on
Store Visits Purchases Conv. Gangnam Stn. 382 79 20.7% Hongdae 298 61 20.5% Jamsil 276 52 18.8% Sinchon 189 33 17.5% * Sample screen · illustrative data Tomorrow · think. Candidates to move, and what was already done.
A screen example. Metrics and thresholds are set to the space.
MODULES & FUNCTIONS
Functions that cross all three modes
Swipe sideways for more · 4
Three technologies behind SAAI
Anonymization
Create data you are allowed to keep
Spatial Intelligence
Turn a space into coordinates you can query
Agentic AI
Propose the next decision on top of it
SAAI
Four letters, one promise
Spatial Anonymized Agentic Intelligence
Anonymized Spatial AI — everything begins on anonymization.
Four foundations, one loop.
- S01Sensesaai careSpatial · Reads the space, beyond the POS.
- 02Analyzesaai insightPuts the same signals into a time series and a pattern.
- A03Thinksaai agentAgentic · Points to what to change, with the evidence.
- 04ActPeople · robots · machinesWhere no product sits.
- I05Verify & LearnFolded into the next judgmentIntelligence · Gets sharper the more you use it.
What acts is not the AI. It is the people and equipment in the space. Steps 04 and 05 are where no product sits.
Three technologies underneath SAAI
Learn more- 03Agentic AIAgentic AI proposes the next operating action.Learn more
- 02Spatial IntelligenceSpatial AI analyzes anonymous behavior signals.Learn more
- 01AnonymizationVideo is anonymized at input (SEAL), and the original is never stored.Learn more
MULTI-STORE OPERATIONS
Many stores. One operating standard.
The three products run on the same data.
However many stores you have, the same metric compares them.
- saai careStore-level issue detection
- saai insightSpace-level operating analysis
- saai agentHQ action history
On the right, branches placed side by side over the same period and the same metric.
View the enterprise packageSame period, same metric, read branch by branch
| Visitor mix | Visits | Zone moves | Avg. dwell |
|---|---|---|---|
| Branch AAll · all ages | 1,576 | 1,672 | 2.9s |
| Branch BFemale · under 30 | 287 | 430 | 2.3s |
| Branch CMale · 30 to 49 | 287 | 239 | 1.9s |
| Branch DAll · 50 and over | 430 | 287 | 5s |
Seven-day basis. What separates the branches is the visitor mix, and every figure comes from the same analysis.
CAPABILITY CATALOG
We have solved a different problem in each space.
The ones nobody has raised yet, we solve together.
saai count
Footfall & occupancy
Counts who passes the door and who walks in.
See detailssaai queue
Queues & crowding
Detects queues and crowding in real time.
See detailssaai ads insight
Exposure and response around in-store media
Compare passersby, approach, attention, and action in one flow.
See details04+
More modules
A problem nobody has raised yet lands here.
* Capability scope and data connections are configured around the customer environment and available data.See how it works and where it is used
saai count
Sense
saai care
Live occupancy and crowding, with alerts when a threshold is crossed
Analyze
saai insight
Daily, weekly and monthly footfall trends by weekday, hour and year over year
Think
saai agent
Staffing, opening hours and promotion timing proposed from forecast demand
USE CASES
Capability choices lead to changes on the floor.
These cases read different questions in different spaces, then examine the next operation under the same conditions.
Convenience · Unmanned stores
Centre-aisle biscuit sales +35%
- BEFORE
- Low back-zone visits · scattered shelves
- AFTER
- Category reset · same-condition remeasurement
Drugstore chain
First shelf by age and gender
- BEFORE
- Experience-led display
- AFTER
- First shelf by customer segment
Shopping mall · large-format retail
Entry inflow and dwell heatmaps
- BEFORE
- One combined entrance total
- AFTER
- Entry-by-entry inflow · dwell zones